(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Four Analytic Studies of the American Literary Braille Code"

FOUR ANALYTIC STUDIES 

OF THE 

AMERICAN LITERARY BRAILLE CODE 



Eric Hamp 
Hilda Caton 
John Siems 



American Printing House for the Blind 

Louisville, Kentucky 
1995 



Digitized by the Internet Archive 

in 2013 



http://archive.org/details/fouranalyticstudOOeric 



FOUR ANALYTIC STUDIES 

OF THE 

AMERICAN LITERARY BRAILLE CODE 



Eric Hamp 
Hilda Caton 
Jolin Siems 



American Printing House for the Blind 

Louisville, Kentucky 
1995 



Table of Contents 



Introduction iii(i)-ix 

A Fresh Look at the Sign System 

of the Braille Code 1 

Toward an Internal Analysis 5 

Problem Areas 6 

Major Steps in Ordering 7 

Braille and Print Writing: Their Mutual Relation. 9 
The Internal Analysis: A New Approach to the 

Braille Code 11 

Summary 16 

References 2 

Toward a Refinement of the Linguistic Analysis 

of American Literary Braille, Grade 2 27 

Background 29 

Toward a Refinement 38 

Procedure 38 

References 47 

An Analysis of Braille Word Lengths 49 

References 83 

The Text Frequency and Incidence of 

Certain Lower-Sign Sequences in American Braille. . . 84 

Categories of Representation . 88 

Brown Corpus Occurrences 93 

Occurrences in Reader ^s Digest 104 

Patterns of Incidence 125 

Relative Frequency 140 

Some Further Conclusions 143 

References 148 

Summary 150 



Introduction 

The monograph which follows comprises four essays which are 
the first fruits of the formal analytic aspect of our research 
prograim on (North) American braille. These four essays are largely 
separate and independent studies, which can be read as defensible, 
integral investigations on their own. They do not concatenate as 
a series with logical implication and essential ordering in their 
structures as, say, four chapters of a book might. Yet they grow 
out of our work in trying to understand the composition, internal 
relations, and functional economy of the braille code; they arose 
as we tried to approach and manage our total task, and thus have an 
ultimate cohesion; they suggested themselves as problems requiring 
a definition and, hopefully, a solution or series of resolutions, 
and they engaged our serious work more or less in the order given. 
For these reasons it seems clear to us that they may usefully be 
presented in the order given. 

We say "more or less in the order given" since these problems, 
and many others which we hope to investigate and report on in the 
future, tended to rush to mind simultaneously or in overlapping 
fashion as soon as we addressed our larger task seriously. It also 
happens that after one task has been begun it must be interrupted 
and another unforeseen problem solved in order to bring the first 
study to an accurate conclusion; there is no need to trouble the 
reader with the details of such necessary housekeeping or 
monitoring. 

The first study is a mild updating, with no substantive 
revision, of Hamp & Caton, JVIB , 78, 210-214 (1984). The notes 



which have been added are intended generally to amplify or clarify 
passages which seem to have offered difficulty to readers and 
associates; the original was written to convey a maximum of the 
analysis reported on in non-specialist language and within brief 
scope in a journal which must husband space wisely. It is clear 
that we were not always sufficiently explicit and felicitous in our 
exposition, and it is hoped that the present reprinting (with 
several typesetting corrigenda) will make this article more useful, 
as well as more available, to those who may be interested in 
following the fate of these braille studies further. 

The purpose of the first study was and is to set forth the 
basis and principles of the analysis of the code (Grade 2, 
Literary) which was in some form prerequisite to the design of the 
elementary teaching materials which were issued as Patterns: The 
Primary Braille Reading Program (1980). It is a very compressed 
account, with the elements identified in the various classes 
presented only in selected illustrative form. In the near future, 
when other pressing desiderata have been disposed of, it is 
intended to issue a more ample presentation and justification of 
this Scime terrain. 

The second study followed the first in written draft at a much 
later date (1989-1991), largely because in the interim we were 
occupied with the planning, writing, revision, testing, and 
production of Patterns: The Primary Braille Spelling and English 
Program (Level A 1992, B 1993, C and D in press and preparation). 
This study announces a plan to expand and refine the analysis 



sketched in the content of the 1984 study. In order to make the 
intellectual background of our investigations and their stage of 
theoretical advancement more readily grasped the development and 
change of 20th century linguistic theory is compactly sketched in 
thumbnail fashion; lest the compression reach caricature 
proportions some references to recent general linguistic 
literature, which may in turn lead to fuller references, have been 
included. 

For our purposes here, the core of the second study is a 
description of the corpus of American English text which forms the 
data base from which our sample has been drawn for the studies here 
described. That corpus is the well known and meticulously 
constructed "Brown Corpus". Its composition, genre constituency, 
and gross quantitative character, and these features of our sample, 
are set forth, as well as our treatment of the text sample for 
entry in computer storage. The study concludes with a sample 
elementary result of processing this text excerpt for a response to 
a simple query: What is the crude count of braille units, classed 
by the categories identified in the first (1984) study, found in 
the sample drawn from the Brown Corpus? 

We say here "crude count" because we rapidly found that if we 
were ever to finish this task in a manageable time it would be 
necessary to pool certain 1984 classes. Meantime a further 
painstaking study has been conducted over the past four years which 
will enable us to give accurate figures and totals for all the 
braille unit classes which we originally desired; and a complete 



report on that portion, together with a revision of the end part of 
this second study, will soon be available. 

The third study, an analysis of word lengths, might seem at 
first blush to be almost tiresomely routine and boring as an 
operation, although the result would certainly be desirable, since 
we really know little about this that has anything to do with the 
linguistic and cognitive processing of language. But it takes only 
brief reflection from the linguistic point of view to recall that 
of all linguistic entities recognized in countless languages by the 
folk intuition of native speakers the notion of "word" is perhaps 
the most elusive: When you hear language spoken there is often no 
stable acoustic mark to match audibly the space you read on paper 
or between braille shapes. Small elements, such as in hit 'em hard, 
cling to others so that it becomes difficult to say what criteria 
distinguish an affix from a clitic (an attached but independent, if 
reduced, word) . Of course, if one seeks semantic criteria, there 
are numerous difficulties in establishing to what degree a word 
should have an isolable meaning (How close to the encyclopedic 
individuality of 'horse' is the meaning of of or that?), whether 
names, e.g. Jacob, truly have meaning, whether two word-like forms 
sharing a meaning are one or two (e.g. prime minister, scarlet 
fever, chaise longue, > lounge vs. input). These, and many other 
riddles or just complexities of reasoning, are problems that worry 
linguists (who must account in detail also for spoken language) 
when they are called on to define and isolate the notion "word" , 
which seems so basic to native speakers . 



But additionally we find when we confront written text that 
there are further difficulties in recognizing word units with 
confidence. Even if we limit ourselves to a simplistic approach 
and accept as words whatever appears between spaces (e.g. counting 
a hyphen as not a space) , there are still conundrums in store for 
us: What do we do with numerals, especially large ones? Are they 
words made of numbers, or phrases composed of numerical words? How 
are abbreviations and acronyms to be regarded? Are mathematical 
formulas and formalized notations a part of normal text? Do 
ellipses exist or are they really their parent full forms? What 
about odd shapes, such as Greek letters or printer's signs? What 
is the rational way to count hyphenated locutions? We have already 
mentioned above the linguistic aspect of difficulty with proper 
names; but there is the further problem of their complexity (e.g. 
New York, San Diego, Baile Atha Cliath, John Smith, Elias Tate III, 
Harry S Truman, Constitution Hall, La Scala, The Reverend Robert 
Walker, and the punctuated phrasal The Atchison, Topeka and Santa 
Fe) and their many idiosyncratic categories {Gone with the Wind, 
She Stoops to Conquer, Xerox) . Allied to the last is the problem 
whether the highly divergent class of foreign words is usefully, 
for our purpose, English. 

Obviously, these problematic inter-space spans have for us an 
interest which is marginal or segregable. Our primary interest 
must be to characterize the English that constitutes the bulk, 
core, and working inventory of essential written discourse of all 
kinds. Therefore, after pondering the categories that seem to 



appear in our sample corpus, we elected to extract these in 
separate lists and to reserve them for separate treatment; we have 
already made extensive progress with the last mentioned, and intend 
to report on that at a later date. At first we feared that in 
following this up we might find ourselves consuming valuable tdLme 
on a tiny corpus of trivial curiosities; as will be seen on a later 
occasion, this corpus-fraction is by no means small or negligible, 
nor is it without rich theoretical and practical interest. 

Having, then, set aside these deviant categories, we have 
reached some interesting results on the length of English braille 
words in relation to inkprint, which we present herewith. It 
should be recalled and emphasized that meaningful results of this 
sort cannot be attained without a prior analysis that identifies 
basic primes of braille comparable to our braille units; hence the 
essentiality of the analysis dealt with in the first (1984) study 
for all these succeeding studies. 

The final study in this volume, one of a number of 
methodologically analogous studies already included in our work 
plans, deals with a feature, or syndrome characteristic, of braille 
structure which is detailed in nature but salient in its formal 
character and important to a serious consideration of braille in 
its social setting: the incidence of certain sequences (of 
inkprint) that by rule do and do not require the braille units 
known as lower-signs. This study, it is believed, illustrates 
sharply the clarity and precision which immediately emerges with 
the application of criteria afforded by a set of primes such as 



7 

those discussed in the 1984 study. On the other hand, since this 
general point has already been suggested, and since an appreciation 
of the highly different behavior of these nine sequences or 
braille-configurations and their actual use will emerge only by 
minimally studying our tabulations, our criteria for 
classification, and our commentary, there is no advantage served by 
further introductory remarks on this study: Those familiar with 
the braille code will already recognize the importance and 
interest — or perhaps the idiosyncratic annoyance — of the "lower- 
sign words". Teachers of braille are familiar with the special 
effort that these handy but quirky items in the tool kit impose. 

When we started on some of these problems over fifteen years 
ago, without access to the kinds of computer-assisted aids to the 
processing of text data which we command today^, we were already 
rapidly aware of the pressing need for precise study of a myriad 
aspects of braille structure, function, and use for the prosecution 
of every phase of the tasks of instruction, for the teaching of 
young and old. We could not have foreseen the gratifying 
possibility that this work might come to be immediately placed in 
relation to fruitful discussions, which now actively make welcome 
progress, leading to a collaborative Unified Code. 

EPH 

15 April 1994 



^But we must never forget that the computer cannot substitute 
for old-fashioned thinking and curiosity, perhaps with the help of 
a brailler or pencil. 



A Fresh Look at the Sign System 
of the Braille Code 

Eric P. Hamp 

Hilda Caton 

American Printing House for the Blind 

1984 
(corrected and lightly revised 1993) 



Abstract 
As one of the steps in the development of Patterns: The Primary 
Braille Reading Program, an analysis of American English braille as 
a written code was undertaken from a linguistic viewpoint. The 
object was to view and analyze the braille code internally and not 
as an encipherment of printed language; thus we view such arrays of 
dots as (brl) as a single meaningful unit for a young blind 
learner, on a par with {and) or g, and not as a "contraction" of 
some full form as yet unknown and forever unseen in its 
presupposedly expanded shape. Such an analysis would be 
appropriate for purposes of instructing those who have never 
mastered or even seen print designed for the sighted. The process 
of analysis was conducted so as to yield the primes which manifest 
the structure of braille and the relations which characterize them. 
The fundcimental relation exploited in order to identify these 
primes was the "sign relation" as notably expounded in linguistics 
by Ferdinand de Saussure. The analysis led to a regrouping, 
identification, and fresh description of the many elements of 
American English literary braille. Grade 2, with the intention to 
facilitate both the literary teaching and the learning of reading 
braille. 



A Fresh Look at the Sign System 

of the Braille Code " ■"'■ 

There is no dearth of documentation to support the assertion 
that much of the difficulty young visually handicapped children 
encounter in learning to read can be ascribed to the fact that the 
materials used have been transcribed into braille from print 
editions . The problems made by this practice have been studied arid 
described in detail (cf. Ashcroft, 1960; Bleiberg, 1970; Caton, 
1979; Lowenfeld, Abel, & Hatlen, 1969; Nolan & Kederis, 1969). 

In an attempt to overcome some of these problems, a primary 
braille reading program has been developed at the American Printing 
House for the Blind. Development of this program. Patterns: The 
Primary Braille Reading Program (Caton, Pester, & Bradley, 1980), 
began with an analysis of all available research related to the 
braille code, including the studies cited above. 

However, as this analysis was conducted, it became apparent 
that the information gained was not adequate for the development of 
the most effective materials for teaching or learning braille 
reading. One major problem was that the existing categories, or 
groupings, of the various elements of the code were confusing to 
both teachers and students — primarily because the descriptive 
language used was inadequate. For example, terms like 
"contraction," "sign," and "symbol," are used interchangeably to 
describe whole-word contractions, part-word contractions, short 
forms, alphabet words, lower-cell words, upper-cell words, dot-five 



words, whole-cell part words, etc. Many other examples of the 
confusion caused by existing groupings and descriptions of the 
braille code could be presented if space permitted;^ but the point 
is that it was not possible to describe the braille code clearly to 
the students who would use it to learn to read. Another problem 
was that no teaching material existed that reflected the internal 
characteristics of the braille code, and teachers were forced to 
use materials reflecting the characteristics of print. That is, 
all sequencing of vocabulary, reading skills, and teaching 
strategies was based entirely on the principles of print reading. 

For these reasons, it was necessary to conduct an internal 
analysis of the braille code that would result in a regrouping and 
fresh description of its various elements, thus facilitating both 
the teaching and the learning of braille reading. The new 
descriptions and groupings allowed teachers to discuss and describe 
the various elements of the code in a clear, precise manner, so 
that their students were not confused. In addition, the analysis 
clarified the internal characteristics of the code, so that 
teachers could devise more effective teaching strategies . 

The authors wish to emphasize that the information provided 
through the research mentioned above is relevant and significant, 
and, in fact, was used to a large extent in sequencing the 
vocabulary and skills in Patterns: The Primary Braille Reading 
Program. We maintain, however, that work before Patterns did not 
sufficiently recognize the truly specific characteristics of the 
braille code; and that the internal analysis discussed in this 



paper does so.^ 

In order to present our results in a clear and concise way, we 
begin by describing the total process involved in the development 
of the reading program; first discussing the problems addressed in 
the design of the course material and in the internal analysis 
itself. 

The entire study was undertaken from a linguistic viewpoint. 
The authors wish to emphasize that the analysis and explication df 
braille throughout the paper is done without appeal to the 
characteristics and setting of sighted print — an intrusion 
irrelevant to our analysis and too often invoked.^ 
Toward An Internal Analysis 

It has long been recognized that a major problem in the 
teaching of braille reading arises from differences between the 
representations of natural spoken language in print and in the 
braille code. These differences have tended to be depicted in 
terms of additions or subtractions; that is, people have been 
occupied with contractions as being somehow shorter by subtraction 
from the full print form, and the like. Attention has also been 
focused on the particularities and tactile difficulties of shapes 
of the dot configurations within cells . This is clearly important 
in itself, yet has little to do with the functions which have been 
allotted to various character groups . 

Less obvious is the fundeimental difference in internal 
structure between many aspects and subsections of the braille code 
and those of a full font of type. A further important point to be 



stressed is that of the nature and status of relations, in an 
abstract sense, between braille and print. That is to say, there 
are clear and partly simple relations that hold between print and 
the braille code — for, after all, the braille code was devised on 
the basis of print, or writing; but frequently the learner of 
braille is not already in possession of a knowledge of print or 
writing . 
Problem Areas 

After reflecting on such questions, we have arrived at the 
conclusion that two large and essential problems confront us at 
this time: 

1 . The present lack of a thorough and relatively abstract 
internal analysis of the braille code stands in the way of our 
ability to confront effectively and confidently many of the major 
decisions involved in designing adequate learning materials. 
Although we do not minimize the high intellectual fascination that 
such a formal analysis holds, we have been impelled to proceed by 
practical concerns; the following analysis reflects this, and the 
ordering of elements is, to a considerable extent, influenced by 
our sense of practicality in teaching and in course design. 

2. In scrutinizing materials and practices employed to date 
in teaching braille, and even those explanatory materials that 
authoritatively present the braille {English Braille American 
Edition 1959), we are impressed by the prevailing tendency to 
analyze the code and explicate problems via elements, processes, 
assumptions, and customs based on visual experience. Even though 



all concerned are well aware of the situation of the blind student, 
and particularly of a young person who has never had the 
opportunity to be exposed to print, much of the discussion and 
explanation is carried on as if all members of the dialogue had a 
degree of experience with the rudiments of printed or written 
English. So, for example, in materials of discussion directed to 
a blind student, the braille unit that conveys the dental spirant 
sound at the beginning of a word such as thin is usually talked 
about as though it were obvious that it should be spelled in 
conventional English with the letters t and h. Of course, no 
teacher of braille thinks of the word thin as starting out with the 
two sounds t and h} and by mentioning this letter combination to a 
person who has never seen it, the teacher can give an initial 
impression only of irrelevancy at best. 

We propose to deal with these problem areas by presenting, as 
compactly as possible, the results of a fresh and completely 
internal analysis of the braille code. 

In this analysis the characteristics of visual print are never 
used as internal elements of the system to be analyzed. Rather, 
the features of visual print are segregated as a separate system 
external to the braille code; that is, as a set of elements with 
which the braille elements may be contrasted, and for which 
relations may later be defined between the two. Our analysis both 
characterizes the internal braille code and states the relations 
between braille and print. We shall return shortly to a more 
precise statement of how this mode of analysis is to be conducted. 



Major Steps in Ordering 

To keep clearly in view our ultimate practical aim in light of 
the problems discerned above, we propose that the goals for the 
devising of teaching materials should follow these major steps: 

1. A relatively complete internal analysis of the braille 
code. We say "relatively complete" in order to isolate major 
issues; to see in which direction one must move; and, finally, so 
as not to overload the analysis and presentation with endles's 
detail or rarely occurring, nonessential elements. We restrict the 
analysis to the generally used major part of literary braille, and 
content ourselves with giving as examples only representative or 
crucially interesting instances for most of the classes discussed. 
A fairly complete list of the resulting categorization is given in 
Box 1. 

2 . Design of the most direct and reasoned route toward 
mastery of the elements and their combinations, and of a compact 
and maximally simple notation to carry out such discussion. This 
means that all teaching materials were ordered for the presentation 
of all elements for which a principled ordering can be determined. 
This is not simply a question of leaving nothing to chance; it is 
a matter of proceeding logically from the known to the less known. 

3 . Introduction of visual elements , including print and 
writing. (As mentioned earlier, these elements must be clearly 
discriminated for the purposes of the analysis; we are not here 
concerned with the problem of the correct introduction of these 
visual elements as a goal of the learning and teaching process.) 



4. A careful consideration to be given, finally, to the 
correct phasing and insertion of the activities specified under 
numbers 2 and 3: in short, the effective teaching of braille and 
the appropriate teaching of print. (Again, this task falls beyond 
our present purpose.) 

As we have just observed, numbers 3 and 4 fall outside the 
scope of the present study. Number 2 depends crucially on number 
1, and the publication Patterns: The Primary Braille Reading 
Program represents the first essay toward fulfilling the objectives 
of number 2. The following discussion, therefore, addresses itself 
to the objectives of number 1 only. 
Braille and Print Writing: Their Mutual Relation 

If no English speaker used any mode of representation for 
English other than braille, the analysis of braille would not 
differ essentially from graphic analyses of one or another writing 
system now or formerly in use in the world. Of course details 
would differ. The writing and printing of English is essentially 
an alphabetic enterprise. The mixed representation that braille 
employs consists of symbols for letters, syllables, and whole 
words. Consider these examples: the letter b in the braille word 
boy, the syllable (er) in the braille word exercise; and the whole 
braille word {people) . An entire world of English speakers using 
only braille would present us with the straightforward task of 
correlating the braille system, internally formulated, with the 
graonmatical and semantic facts of the English language. Thus this 
job would be analogous to, and only more complex than, that of 



correlating conventional English spelling with the way our words 
are put together. But braille users do not constitute the entire 
English speaking population; nor is it expected that those who use 
braille will fail to master conventional print English also. 

This fact imposes a separate dimension on the analysis, with 
a corresponding task for analyst and teacher; a dimension arising 
from two principal facts of language use that have impinged on the 
character and function of braille. 

1. Historically, of course, braille is not independent of 
print English. It has been devised over time, with print English 
preceding chronologically and lurking in the background as a 
partial model. 

2. Braille will, moreover, be learned by students who wish to 
convert with maximum efficiency that knowledge and skill into a 
competence with English print, so that they can type and of course 
use a computer. 

If our analysis of braille is to take these two aspects into 
consideration, the resulting formulation of the relations of the 
internal braille system with the structure of print English will be 
complex in a special way; that is, it will distinguish more classes 
of elements and functions than if only the braille system or only 
the structure of print were being considered. 

At every stage of this analytical process, we must ask 
ourselves: 

1. What forms are distinguished by what elements and what 
combinations in terms of braille shapes and braille units . 



10 



2. How each of these forms is correlated with, that is, 
equivalent to, not equivalent to, or partially equivalent to, 
elements and features and combinations of print English. 

We see immediately from this that there are two important 
aspects to every determination that we make for a braille element: 
the configuration of the element itself (so and so many cells 
consisting of such and such arrangements of dots), and the element 
or elements of print English with which this braille shape(s) is 
found to be correlated. , - . 

In the theory of signs as elaborated in the technical 
literature of linguistics, such a relation is known as a "sign 
relation," and the combination of, or more technically, the linkage 
between the signifier (the braille configuration in this case) and 
the signified (the print English) is known as a sign.* To invoke 
an analogue, this value, or function, stands in relation to the 
signifier much as a meaning does in relation to its signifier which 
we call a word. Thus, what we are calling value here is much akin 
to the notion of meaning} but, for the present, we shall not use 
the latter term in this sense, lest we cause confusion with a 
different type of sign relation; naunely, that of linguistic 
semantics, with which we are not here concerned. 

This consistent correlation of print values with the 
distinguished braille elements may be called for convenience a 
value system, or sign system, in the sense just specified. It is 
in such an analysis, as well as in the details of its execution, 
that our approach differs fundamentally and to the greatest extent 



11 



from earlier formulations and presentations of the braille code. 

The following internal analysis presents a relatively detailed 
description of this consistent correlation of print values with the 
particular braille elements corresponding to them. The 
presentation consists of a set of new terms and the regrouping of 
all elements of English braille, Grade 2, as well as the new 
descriptions of these elements previously discussed. 
The Internal Analysis: A New Approach to the Braille Code 

To begin our internal analysis, it is necessary first to 
distinguish the following primes of braille. For clarity, we have 
restricted the meaning of each prime to a single sense. 

Cell . This term has been used in more than one sense in the 
literature. Here cell is defined as an abstract space twice as 
high as it is wide, in which there are six positions arranged in 
three rows and two columns, in which dots may appear. 

Shape, We defined as a shape a single configuration made up 
of one to six dots and occupying a single cell. Note that, so long 
as a shape is defined in this fashion, it does not yet have any 
necessary value or meaning, and therefore is not a sign in the 
sense used in the sign theory alluded to above. 

Dot . A dot is defined as the element of which shapes in a 
cell are composed. The dots of a braille shape occur physically as 
bumps or bosses. 

Braille unit. We now require a term for any shape or shapes 
taken together in correlation with its (or their) value or meaning. 
It would be convenient and nonoffensive to ordinary English if the 



12 



word character could be used for this concept, especially since 
that is the established English term for the element of Chinese 
writing that closely resembles in function this braille element. 
But unfortunately this term has already been employed for some time 
by users of braille in a deceptively similar but ambiguous sense; 
and so, unless users of braille wish to alter their terminology, it 
will be necessary to find some other term for this pivotal concept 
in the present analysis. For this purpose the term braille unit 
has been proposed by us.^ 

A single braille unit may consist of one or more shapes; as, 
for example, (go) = 1 shape, (tion) = 2 shapes, etc. Braille units 
fall into three major types — letters, modulations, and grams. 

1. Letters. Letters are either alphabetic, 
such as jb, or nonalphabetic , such as 2 . 

a. Alphabetic letters (or letters proper) have print- 
alphabetic values; that is to say, these braille units match 
occurrences of ordinary letters in print. For example, the shape 

• • 
( T ! ) has the value of i . 

b. Nonalphabetic letters comprise (1) numbers, the 
decimal point, and the fraction bar, and (2) certain other braille 
units with abstract letterlike segmental function, such as the 
asterisk and the apostrophe. The numbers, of course, take the 
nujjiber sign, and then the shapes with the resulting number value 
may be thought of as letters of a numerical alphabet consisting of 
twelve shapes (counting the decimal and the fraction bar) and 



13 



i i 



spelling number words . The asterisk may be viewed as a rather odd 
unpronounceable letter, and the apostrophe often has the implied 
meaning "letter left out." The reason for classing these two as if 
they were letters or numbers is that, like conventional letters, 
they have sequential segmental properties in linear order, and the 
characters they match in the print text consist regularly of single 
segmental print shapes . 

2. Modulations, Modulations are of two rather different 
sorts: punctuation, as for example, the question mark; and 
register, as for example, italics. What these have in common is 
that they "do things" to, that is, have effects on, other elements 
— the segmental elements — in the chain. 

a. Punctuation, in fact, has print values that are, 
themselves, sequential in position; but these braille units differ 
from others in having domains of effect extending at times to 
considerable distances to the right and left of their sequential 
position. Some punctuation looks back; examples are the period and 
the non-Spanish exclamation point. Other punctuation encloses; 
excunples are the hyphen and the dash. Those that look back have as 
a domain of their force what has gone before; those that enclose 
both warn us of their application and close their domain; those 
that link affect things on both sides. 

b. jRegister is the term applied to those braille units 
that include what have traditionally been called composition signs} 
these units look forward, and may also automatically specify where 
the scope or domain terminates. Examples are capital, italic. 



14 



letter and number signs. These elements always have the effect of 
modifying the basic segmental values of what follows; thus, they 
change the dress of some elements, such as lower case into 
capitals, or change what we think of as type style, such as italic, 
or change letters into numbers, or change the abstract reading of 
an element, such as the letter sign. Registers have the unique 
property of finding no separate segmental counterpart in print. 

3 . Grams . These are of three kinds : phonograms , such as the 
(ance) in dance; morphograms, such as the (ance) in reliance} and 
logograims, such as the words (rather), (the), {friend) , (mother), 
and (immediate) . The distinction between grams and the two 
preceding units is that, unlike modulations, they are segmental in 
value; but, unlike letters, they have no single counterparts in a 
type font. Because of this last property, these are the braille 
units that later will give rise to bidirectional problems in 
writing and spelling. 

a. A phonogram is a braille unit having a phonetic value 
that would be written in print by more than one alphabetic symbol. 
Phonogrcuns are either one-shaped, like (th) , (ch) , (gh) , the (ing) 
in sing, the (ea) in read, the (ed) in bed, and the (ar) in bar, or 
multishaped like the (ation) in nation, the (ound) in sound, the 
(ong) in long, the (ence) in fence, the (ity) in pity, the (ness) 
in Tennessee , and the (less) in Jbless. 

b. A morphogram is a braille unit having the value of an 
element in a word, such as an inflectional ending, prefix, or 
suffix. Examples are the s in words, the (ing) in looking, the 



15 



(ed) in looked, the {ance) in avoidance, the {ation) in admiration, 
and the (in) in inconsistent . Note that the shape(s) that make up 
(ing) , [ed) , {ance), etc. appear as phonograms or morphograms, 
depending on their function in words, that is, their "value." 

c . A logogram is a braille unit made up of one or more 
shapes having the value of an English word (conventionally, a chain 
of letters between spaces) with either limited reflection or no 
reflection at all of phonetic values. There are two principal 
configurations of logograms, single-shape and multishape. Single 
shapes comprise letter words and wordlets. A letter word is a 
logogram that has a shape the scime as that of a letter. Examples 
are (but), {can), {do), {rather), Wordlets comprise all other 
logogreons carrying a word value. Examples are {and), {the), 
{shall), {still), {there), {ought), {young), {those), {enough), 
{cannot), {paid), {declare), {was), {to). It should be observed 
that logograms do not lose their status as such when they undergo 
derivation by affixes. Thus {spirit) remains a logogram, in this 
case a compound-letter word (let), when it occurs as part of the 
derived adjective spiritual. Such a definitional provision avoids 
the need for pedantically encumbering the analysis by renaming 
hosts of wordlets "morphogrsuns" because they may form parts of long 
derived words.* 
Summary 

As we have stated, the internal analysis as described in this 
paper was used for regrouping the various elements of the braille 
code and devising new descriptions of these elements. This was 



16 



done to eliminate the many conflicting and confusing terms and 
categories previously used by teachers and students, and to provide 
them with a new system consisting of a relatively small number of 
categories, or groups, and clear, precise, linguistically based 
descriptions of all the elements of English braille. Grade 2. 

In addition, we have presented a discussion of the braille 
code, based on linguistic principles, which is intended to clarify 
the internal characteristics of the code. Specifically, it is 
intended to dispel the notion that the braille code and the 
principles of braille reading are in a way analogous to the print 
code and the principles of print reading. 

We believe that the results of this analysis can be used to: 

1. Provide teachers with an outline of braille terms that 
will enable them to describe and discuss any element of the braille 
code in a manner easily understood by children. ^ 

2 . Emphasize that the teaching of braille reading and print 
reading are not analogous, and promote an understanding of the 
internal characteristics of the braille code, thus providing 
teachers with more effective strategies for teaching reading to 
children who use braille as their primary medium. 



17 



Notes 
^In fact, the analysis here presented resulted directly from the 
efforts on the part of the first-named author to learn the braille 
code from scratch by using the handbook (EBAE, 1959) and from his 
need to clarify and restate the foinnulations and expositions set 
forth in that influential work. 

^We claim therefore that earlier work was not merely inefficient 
and inadequate in its obligation to the blind learner, but that it 
was non-descriptive of braille. 

^Of course, because both braille and the inkprint some of you are 
reading are equally representations of the same language and 
because the designers of braille patterned it on inkprint it is not 
accidental that the two have much in common. We therefore do not 
IGNORE inkprint; quite the reverse. The issue is rather where we 
must seek our analytic criteria, and to what relations and status 
we allot all elements . 

""The authorative source on this concept is [de] Saussure (1916). 
^Brunit was considered but did not find favor. 

*A similar example to spiritual would be worldly. Some readers may 
wonder about a specimen such as (some). Would the sequence (soine) 
in lonesome be similarly a wordlet? The answer is no. The 
analysis will have already isolated and identified the suffix 
{some) in {lonesome) , {handsome) , etc. (even though this ancient 
and poorly productive English suffix is not very frequent or 
salient today) ; thus the braille unit {some) here is a morphogram. 
It is not the same as the logogreim {some) . The fact that these two 



18 



have radically different meanings is, however, a linguistic fact, 
but not directly a feature of the status of these braille units in 
the code. 



19 



References 
Ashcroft, S. C. (1960). Errors in oral reading of braille at 

elementary grade levels. Unpublished doctoral dissertation. 

University of Illinois. 
Bleiberg, R. (1970). Is there a need for specially designed 

reading series for beginning blind readers? New Outlook for 

the Blind, 64, 135-138. 
Cardinale, J., & Cline. C. (1973). Methods and procedures of 
V, braille reading. AFB Research Bulletin, 171-183. 
Caton, H. R. (1970). Institute report: Primary braille reading. 

Louisville, KY: American Printing House for the Blind. 
Caton, H. R. (1979). A primary reading program for beginning 

braille readers. Journal of Visual Impairment & Blindness, 

73, 309-313. 
Caton, H. R. , Pester, E. J., & Bradley, E. J. (1980). Patterns: 

The primary braille reading program. Louisville, KY: 

American Printing House for the Blind. 
Chall, J. S. (1967). Learning to read: The great debate. New 

York: McGraw-Hill. 
Davies, A. M. (1991). International Encyclopedia of Linguistics . 

Oxford, 
de Saussure, F. (1916). Cours de linguistique generale, Lausanne 

— Paris: Payot, 1916 [Preface signed Geneve, juillet 1915], 

2nd ed. Paris: Payot, 1922 [reviewed by L. Bloomfield, Modern 

Language Journal 8 (1923): 317-19=no. B17, Bloomfield 

Anthology 1970=1987 abridged ed. , pp. 63-5], 3rd. ed, Paris: 



20 



Payot, 1931, unchanged "4th" ed. Paris: Payot, 1949 (reprint 
1955), "5th" ed. 1960 (reprinted since); the entire book 
synthesized and reconstituted by Charles Bally and Albert 
Sechehaye, with the collaboration of Albert Riedlinger, from 
contributed students' notes posthumously assenibled from FS's 
three lecture courses 1906-11; many translations since the 
Japanese (Tokyo 1928), of which in English tr. Wade Baskin, 
Course in General Linguistics, New York: Philosophical 
Library, 1959 (London 1960), McGraw-Hill, 1966 [These 
printings are often referred to by page], rev. ed. (with 
introduction by J. Culler) London: Fontana, 1974 [see also J. 
Culler, F.de S., Ithaca NY: Cornell University Press, 1976; 
rev. ed. 1986]; there are also less satisfactory translations 
and commentaries; basic text edition: R. Godel, Les sources 
manuscrites du Cours de linguistlque generale de F. de 
Saussure, Geneve: Droz/Paris: Minard, 1957; critical edition 
with textual variants: Rudolf Engler, Ferdinand de Saussure 
Cours de linguistlque generale , Wiesbaden: Harrassowitz 1967- 
68 (3 fascicules) + 1974 (4th fasc); critical ed. with tr. 
and commentary originally in Italian (1967, 2nd ed. 1968): 
Tullio de Mauro, Edition critique du Cours de linguistlque 
generale de F, de Saussure, Paris: Payot, 1972; ed. by Eisuke 
Komatsu and tr. by Roy Harris, Saussure^ s Third Course of 
Lectures on General Linguistics (1910-1911) From the Notebooks 
of Emile Constantin, Oxford etc.: Pergamon, 1993. The 
literature on Saussure is vast; see for example, E. F. Konrad 



21 



„, ^ 



Koerner, Bibliographia Saussureana 1870-1970 , Metuchen NJ: 
Scarecrow Press 1972. 

English Braille American Edition. (1959). American Printing House 
for the Blind, Louisville, KY: 1972. 

Lowenfeld, B., Abel, G. L., & Hatlen, P. H. (1969). Blind 
children learn to read. Springfield, IL: Charles C. Thomas. 

Nolan, C. Y., & Kederis, C. J. (1969). Perceptual factors in 
braille word recognition. American Foundation for the Blind. 
Research Series No. 20, New York: American Foundation for the 
Blind. 

Pedersen, Holger. (1924). Sprogvidenskaben i det nittende 
aarhundrede: Metoder og resultater. Kjobenhavn, Denmark: 
Gyldendalske Boghandel Nordisk Forlag; translated Spargo, J. 
(1931). Linguistic Science in the nineteenth century. 
Cambridge, MA: Harvard; reprinted with changed title: 
(1962). The discovery of language. Bloomington: Indiana 
University Press (Midland Book) [authoritative classic by a 
magisterial participant; see pg. 289 of translation, 266 of 
original, for picture of F. de Saussure; and pg. 63 of 
translation for picture of and reference to work of Pedersen 
who did not include these in his original.] 

Saussure: See de Saussure. 



22 



Table 1 

Outline of Braille Terms — Examples 
Letters 

a. Alphcdaetic letters (26) 

b. Nonalphabetic letters (12) - ^ 

(1) 0-9 
decimal point 
fraction bar 

(2) Other braille units with cdastract letter like function 
accent sign 

apostrophe '■ 

asterisk ..^ 

ellipsis 

hyphen or dash — when used to indicate missing letters in words 
Grams 



a. 



Phonograms 








{ally) 


Sally 


(ed) 


red 


(ance) 


dance 


(en) 


pen 


{and) 


sand 


( ence ) 


fence 


(ar) 


car 


(er) 


certain 


{ation) 


nation 


(ever) 


several 


{bb) 


rubber 


iff) 


duffle 


(Jble) 


table 


{for) 


forest 


(cc) 


occur 


{ful) 


awful 


(ch) 


chair 


(99) 


suggest 


{com) 


come 


(9h) 


ghost 


{con) 


contrary 


{here) 


adhere 


{dd) 


paddle 


{in) 


pin 


{dis) 


dispel 


{ing) 


sing 


(ea) 


read 


(ity) 


city 



23 



Morphograms 

{after) afterlife 



{ally) 


mathematically 


{ance) 


avoidance 


{and) 


multiplicand 


(ar) 


secular 


{ation) 


admiration 


{be) 


befriend 


{com) 


commiserate 


{con) 


confuse 


(dis) 


disengage 


(ed) 


rubbed 


(en) 


encephalogram 


( ence ) 


providence 


(er) 


zipper 


Logogram 




(1) Letter 


word 


{but) 


{knowledge) 


{can) 


{like) 


{do) 


{more) 


{every) 


{not) 


{ from ) 


{people) 


(go) 


{quite) 


{have) 


{rather) 


{just) 


{SO) 



{ful) 


wonderful 


{here) 


cohere 


{in) 


indecent 


{ing) 


singing 


Uty) 


rarity 


{less) 


useless 


{ment ) 


ornament, monument 


{ness) 


openness, oneness 


{ sion ) 


aversion, confusion 


{ some ) 


loathsome 


{ through ) 


throughout , throughway 


{ tion ) 


reaction, prediction 


{th) 


seventh 


{there) 


therefore 



{ that ) 

(US) 

{very) 

{will) 

{it) 

{you) 

{as) 



24 



( 2 ) wordlet 
(about) 
( above ) 
{according) 
{across) 
{after) 
{afternoon) 
{afterward) 
{again) 
{against) 
{almost) 
{already) 
{also) 
{although) 
{altogether) {by) 
3 . Modulations 

a . Punc-tuation 

(1) look back 
colon 
comma 
exclamation point 

(2) enclose 

bracket or brace (in pairs) 
comma (in pairs) 
parenthesis (in pairs) 
quotation marks, single (in pairs) 
quotation marks, double (in pairs) 

( 3 ) link 

bar long dash 

hyphen bracket or brace (one) 

dash 



{always) 


{ cannot ) 


{ever) 


{know) 


{and) 


{character) 


{father) 


{herself) 


{be) 


{child) 


{first) 


{himself) 


{because) 


{ children ) 


{for) 


{itself) 


{before) 


{ conceive ) 


{friend) 


{thyself) 


{behind) 


{conceiving) 


{good) 


{myself) 


{below) 


{could) 


{ great ) 


{yourself) 


{beneath) 


(day) 


{had) 


{oneself) 


{beside) 


{deceive) 


{here) 


{ourselves) 


{between) 


{deceiving) 


{him) 


{ themselves ) 


{beyond) 


{declare) 


{his) 


{yourselves) 


{blind) 


{declaring) 


{immediate) 




{braille) 


{either) 


{in) 




{by) 


{ enough ) 


{its) 





period 

question mark 
semicolon 



25 



b . Register 

capital sign, single letter sign 
capital sign, double number sign 
italic sign, single termination sign 
italic sign, double 



26 



[. 



Toward A Refinement of the Linguistic Analysis 
of American Literary Braille, Grade 2 

Hilda Caton 

Eric P. Hamp 

Braille Research Center 

American Printing House for the Blind 

1991 



27 



Abstract 
A plan is broached to continue and refine the analysis outlined in 
Hconp and Caton, 1984. This plan has been pursued since 1989. The 
conceptual foundations of the analysis are placed in the context of 
20th century linguistic theory in brief and summary form. To 
pursue such a plan the first requisite is a suitable text corpus to 
serve as an empirical data base. The creation of such a corpus is 
described. As a scimple result we offer classed totals of braille 
units contained in our text sample. 



28 



A Refinement of the Linguistic Analysis 
of American Literary Braille, Grade 2 

Background 

As one of the components employed in the development of 
Patterns: The Primary Braille Reading Program, an analysis of 
American Literary Braille, Grade 2, was performed drawing on the 
model of linguistic analysis. Results of that analysis were used 
in designing materials and recommendations to serve the teacher of 
young children who will use braille as their learning medium. That 
analysis provided teachers with a means of identifying the objects 
of study, of simplifying explanations of some of the unique 
characteristics of the braille code, and also provided a basis for 
ordering the presentation of braille units to young children. A 
summary and non-technical report of this analysis was published in 
the Journal of Visual Impairment & Blindness (Hamp and Caton, May 
1984), and is here reissued in the preceding pages. That report 
itself is brief and was intended only to indicate the general 
directions taken. The analysis proved to be useful for the stated 
purposes, but was never considered by the authors to be exhaustive 
or complete. For this reason, the purpose of the research program 
outlined in the following pages will be to refine and expand the 
initial analysis. Before some directions of the refinement and 
expansion are summarized, however, it is necessary to present as 
background a brief historically arranged summary of the main lines 
of development of linguistic theories — specifically theories of 



29 



grammar — which directly affect this braille analysis. The 
presentation which follows is greatly simplified and compressed. 
It is intended only to indicate and call to mind in summary the 
rapid movement which the field has witnessed since World War I. 
For relatively accessible introductory accounts and for references 
^■-' - toward further reading one may consult Crane, Yeager, & Whitman 
(1981), Akmajian, Demers, (Farmer), & Harnish (1979, 1990), Finegan 
& Besnier (1989), O'Grady & Dobrovolsky (1989), and always with 
profit Lyons (1968) . As a mode of approach, the following question 
is posed, and aspects of the answers are given: 

Where did linguistic grammatical theorizing and analysis of 
the braille code stand at the beginning of the 1990s? 

A. We will set aside the historical and genetic comparative 
^ linguistic theorizing of the 19th century (and later up 

to the present day) as not being relevant to braille 

problems at all. On this subject see Pedersen (1924, 

""■' 1931), Morpurgo Davies (1991). Between the 1920s and 

the 1950s a group of theories of grammar was proposed 

' ^ - • which is often called "structural." An important aspect 

" •' of that point of view was the insistence on isolating, 

• "" identifying, and characterizing fundcimental entities 

'^" ' separately and specifically relevant to the structure of 

each different language. Different subvarieties of such 
theorizing arrived at, or emphasized, differing criteria, 
and advocated different solutions of detail; but the need 
to recognize relevant and principled basic entities was 



30 



generally agreed. The entities recognized by traditional 
grammar — words, roots, affixes, sentences, etc. — had been 
customarily defined or specified by inconsistent, 
overlapping, or non-empirical criteria. The entities now 
recognized usually comprised superficial components that 
occur as sequences one after the other to make up audible 
spoken or readable written sentences of language; they 
were preferably not abstract objects or relations or 
categories that fail to surface as observable segments or 
spans in such sequences . An important shortcoming of 
this brand of theorizing was the visible inability to 
incorporate such analytic formulations in a satisfactory 
view of the function and variation of such structures in 
the use of human language. Theorizing since the 1950s 
has concentrated on repairing that shortcoming in our 
perception of the goals of grammar. 
B. A major revision in grammatical theory came towards the 
end of the 1950s with what has since been generally 
called "transformational grammar." One important 
consequence of this theoretical view is the realization 
that a correct recognition of the full functional (or 
"semantic" or perhaps "semiotic") properties of a grammar 
will not necessarily preserve at every stage of analysis 
all isolable grsimmatical entities in a constant state of 
relevance to the analyst, nor in the same apparent 
categorization. In other words, just because you can 



31 



identify an element analytically, this does not mean that 
every such element is equally pertinent, or pertinent to 
the Scime degree, to all aspects or phases of the grammar. 
So also the following six equivalent sentences: 

John made the cabinet. 

John was the maker of the cabinet. 

John was the cabinet's maker. 

It was John who made the cabinet, 
and perhaps 

The cabinet was made by John. 

The making of the cabinet was John ' s . 
recast the elements into quite different syntactic roles, 
while scarcely altering the semantic values. What 
changes here is mainly the stylistic, discourse, 
pragmatic, register, or elocutionary value. But this 
realization and a successful formulation of the rules for 
constructing such equivalences do not eliminate the need 
to have agreed and well founded ways for recognizing and 
segregating perceived elements. 
C. Earlier work on transformational grammar (for which the 
term "generative" has often been used) during the 1960s 
introduced many changes, both of detail and of substance, 
in the theory, and even led to a rejection of the view 
concerning which cognitive phenomena are basic to a 
grcimmar (e.g., the doctrines called "case grammar" and 
"generative semantics"). During the 1970s and 1980s a 



32 



sizeable number of competing or variant theories of 
" grammar has additionally been developed, presented and 
debated. These include those known under the names or 
doctrines of "extended standard theory," "trace theory," 
"pragmatics," "discourse analysis, " " functional grammar , " 
"relational grammar," "Montague grammar," "arc-pair 
grammar," "space grammar," "government-binding," 
"generalized phrase structure," "natural phonology," 
"metrical phonology, " "auto segmental phonology, " "lexical 
morphology," "template morphology," not that we attempt 
to exhaust the list. 
D. It will be seen that there are at present among linguists 
many different points of view as to precisely how an 
optimal grammar is internally constituted. All of this 
is really quite independent of the fact that modern 
linguists actually agree on a surprising number of basic 
features, aspects, properties, functions, and phenomena 
that can be observed and stated for any human language 
thus far investigated with care. All of this question of 
internal formulation is independent of the different 
emphases and vantage points from which language may be 
viewed by scholars of particular domains; such scholars 
are singled out or categorized as sociolinguists, 
cognitive scientists, students of perception and 
phonetics, dialectologists, comparative and historical 
linguists, theorists of spoken versus written language. 



33 



on which latter see Senner (1989), not to mention those 
specialists known as anthropologists, ordinary language 
philosophers, logicians, specialists in artificial, 
computer, and restricted languages, semanticists, 
students of data retrieval and artificial intelligence — 
all specialists in some obvious disciplines that exploit 
or impinge on linguistic analysis and theories which may 
be called grcunmatical . 

E. The goal of grammarians in the 1940s, 1950s, and 1960s 
was to write a complete grammar for every language. With 
the later discussions of grammatical theory and the 
resulting realization of the great complexity of 
language, taken in its most inclusive sense, that goal as 
a literal possibility has been largely given up in any 
short term understanding of a reasonable target. 
Instead, today, grammarians mainly write or frame 
paradigmatic discussions of interesting, crucial, or 
boundary-defining areas of the arguments with the hope of 
arriving at an incisive definition of the problem; and at 
a pertinent and explanatory formulation of a regularity, 
or rule of greunmar, or of the coverage and function of a 
portion of the assumed grammar. 

F. It is therefore seen that the modern state of linguistics 
does not claim to discover definitive grammars in a final 
way. Rather, it seeks to work towards the refinement of 
grammatical questions, at the same time clarifying 



34 



grammatical theory. 

G. It is therefore at present the case that formal 
grammatical descriptions are presented to touch upon the 
total range of a language only in the case of languages 
and dialects which have never really been investigated at 
close range in a modern, or 20th century linguistic mode; 
and the model for presentation in such cases is generally 
that of the sort of observation and analysis referred to 
under A above, with excerpts from B or C. At some later 
time, or even by anticipation along with this first 
analysis, an analytic study in one of the frameworks of 
B, C, and D above may also be carried out. The natural 
languages which are now analyzed in this way tend to be 
those many languages of the world spoken by smaller 
populations and often called "indigenous." 
In such a freimework many concerned linguists today 
conduct salvage work on endangered languages which 
tragically find themselves on the brink of extinction. 
Such studies are quite separate from the topic of 
investigation which has been dubbed "language death" or 
"obsolescence. " 

H. In preparation for the development of Patterns: The 
Primary Braille Reading Program, an analysis of braille 
was carried out by Hamp essentially within the 
theoretical framework outlined in A above, or an analog 
thereof. In this sense the presently attained analysis 



35 



of braille is at the theoretical level of linguistics in 
the 1950s, just a third of a century out of date. 
I . The analysis of braille was not carried to any of the 
further above modes at that time because: 

1. our manpower and available time were limited; 

2 . it was essential to keep matters simple and 
manageable and not to become engrossed in recent 
theoretical debate, so as not to lose our goal from 
view; 

3. it seemed possible to reach a useful result with 
that mode's investment of theory; 

4. it was not certain, for a start, that a profession 
anchored in tradition would rapidly accept the 
radical changes entailed even by our modest 
proposal shorn of forbidding complexities drawn 
from technical linguistic discourse; 

5 . it was and is not clear that certain given modes of 
more probing analysis will equally give immediate 
yield for our grasp of braille as a limited object. 

6 . of course it will be ultimately important to 
understand the properties of braille in their full 
extent and context as a symbolic system, and we 
hope to participate in the exploration of this as 
it unfolds in due time. 

J. From the reception accorded our first set of efforts, we 
are encouraged to believe that further development in 



36 



both the adaptation of theory and the accomplishment of 
analysis will lead to formulations and descriptions of 
portions and aspects of braille writing which will in 
some measure mirror the positive developments of theory 
and analysis reflected for the field of linguistics in 
points B, C, and D above. For the time being, because of 
the priority of duties in our profession which demand 
productive results all along the way, we defer 
exploration of the rather more theoretical of these aims . 
Our present tasks are much more like those of a 
composition class in a village school than of a syllabus 
for a university linguistics course. But we insist that 
in our analytic duties we must keep the later goals and 
lesson in view. ;;;;; .-^ ? - : 
K. In the meantime, we perceive a pressing set of tasks 
which should be addressed. These tasks remain 
theoretically within the framework of points A and B 
above; they are designed to build upon, consolidate, 
refine, and further select for concentrated application 
the admittedly rather crude results that were obtained in 
our first analysis and presented for nontechnical reading 
in the outline sketch (Hamp and Caton, 1984) which is 
reproduced above. Those results, however, have proved 
fundamental and essential, if rudimentary, to our further 
work in the interim. 



37 



Toward a Refinement 

It is clear that for the empirical study of English braille as 
used in North America an acceptable corpus of text must be selected 
and appropriately prepared. This task has therefore been our first 
priority over the years 1989-1993. Though some results have been 
rapidly obtained, we still have elements of detail to polish and 
report . 
Procedure 

The text materials used for the present analysis are 25 
samples chosen from the corpus which forms the basis of the 
publication. Computational analysis of present-day American English 
(KuCera, H. , & Francis, W. N., 1967), generally known as the "Brown 
Corpus." This corpus comprises 1,014,312 words of prose and 
consists of 500 samples of about 2,000 words each, taken from 
contemporary publications in American English. The seunples include 
scientific and learned writing as well as fiction, journalistic 
prose, and other genres; the character of the samples was carefully 
controlled and explicitly specified. A meticulous characterization 
and listing are to be found in KuCera and Francis 1967 and in their 
manual (revised and amplified 1979). The 25 samples used in the 
present analysis were selected so that the text materials may be 
representative of all types of literature included in the "Brown 
Corpus." The proportionality and inventory of the samples drawn 
from the Brown Corpus are as follows, with genre indicated in each 
instance. 



38 



Genre 



Number of texts in corpora 
Brown Ours 



Informative; 

A. Press: Reportage 

B. Press: Editorial 

C. Press: Reviews 

D. Religion 

E. Skills and Hobbies 

F. Popular Lore 

G. Belles Letters, 
Biography, etc. 

H. Miscellaneous 

J. Learned & Scientific 

Imaginative ; 
K. Fiction: General 
L. Fiction: Mystery 

and Detective 
M. Fiction: Science 
N . Fiction : Adventure 

and Western 
P. Fiction: Romance, Love 
R . Humor 



44 


2 


27 


1 


17 


1 


17 


1 


36 


2 


48 


2 



75 4 

30 1 1/2-2* 
80 4 



29 



1 1/2-2 



Our Samnles 


Al, 


A21 


. Bl 




CI 




Dl 




El, 


E21' 


Fl, 


F21 


Gl, 


G21, 


G41 


, G61 


Kl, 


K21 


Jl, 


J21, 


J41 


, J6 


Kl, 


K21 





24 


1 


LI 




6 


> ; 


- ■ . 




29 


1 1/2-1 


m 


ove 


29 


1 1/2-1 


ti 




9 


1/2-1 


Ri 


Total 


500 


25 





♦These are rounded at risk of bias 



39 



The English texts which were selected to be used as a basis 
for the analysis were copied from the full "Brown Corpus," and a 
new file was created containing 25 samples selected from those 500 
texts which constitute the corpus . 

Data of the full text had been scanned to identify the visual 
character set and symbolism used in the version of the text which 
had been received. 

The text of the file of selected samples was edited for 
translation into braille. Some of the steps in editing were: 
distinguishing between opening and closing quotes where the quote 
character (") appeared; distinguishing single quotes from 
apostrophes, both indicated by (') in the text; marking the 
headings for appropriate braille representation; indicating which 
letters standing alone required letter signs; editing for italics; 
and identifying foreign words and acronyms in which braille 
contractions should not be used. 

After editing, the samples were translated into braille and 
proofread by a braillist. 

In the output from the braille translation progreon, each 
braille cell is represented by a four-digit number. Part of this 
number indicates the dots which form the shape. This part of the 
niimber provides the data for embossing. The other part of the 
number serves to distinguish between two cells (often referred to 
as "signs") with different meanings regardless of their possible 
identity in shape. For instance, {gg) and (were) have code numbers 
of which three digits are the same and one is different. This 



40 



unique identification is used for checking the accuracy of the 
program. 

The four-digit cell numbers were grouped to correspond to 
braille units. This resulted in a data set in which (not), {ness) , 
and (neither) , as examples, were represented by four-, eight-, and 
twelve-digit numbers respectively. 

A counting program was developed having a table of the braille 
units. Each entry contained the numeric value of the unit, the 
category of the unit, an inkprint equivalent, and a unit naune. The 
names were limited to four characters using an abbreviated form 
suggested by the braille in some cases, as Roman "mch" in 
representing {much), and arbitrary abbreviations in others, as 
Roman "whch" for {which) . Thus it should be clearly understood 
that these representations are purely arbitrary tags of 
convenience. 

The program reads the file of braille unit numbers and prints 
rows of dots representing the braille cells, a line beneath the 
dots showing the inkprint equivalent, and, between these, a line 
marking the beginning of each braille unit . 



41 



• • 



dP M 



H 
cN9 Eh 



CN 

I 
I 



Hi 
Eh 
O 
EH 



I 
I 
I 



O 



• • • 

• •• 

• • • 

• • • 



• • • 

• • • 



<ifi Eh 
e¥» X 

dP Eh 
I S 

o 



1 


1 


1 


1 


1 
1 


1 
1 


EH 


o 


Q 


CO 




Pi 


O 




pa 


tf 




PM 



CN 



.1 t 



• •• 

• • • 



• • • 

• •• 



1 


1 


1 


1 


1 


1 
1 


1 

1 


1 
1 


J 


S 


>H 


CO 

H 
» 
Eh 



0) 

•H 
fa 



• • • 

• • • 

• •• 



• • • 

• •• 

• • • 

• • • 



vo 



CM 



1 1 
1 1 


1 
1 


1 
1 


1 1 


1 


1 


§ *^ 


X 





• • • 

• •• 



cW> Eh 



I Z 



dP Eh 
eW> pq 

df> >4 



1 
1 


1 
1 


1 
1 


1 
1 


1 


1 


1 


1 


a; 


H 


E^ 


z 


» 






H 


E^ 








o 









I 
I 

z 
o 

z 



n 

I 
I 
I 



CO 



I 
I 
I 

Z 

o 
u 



a\ 



• • • 

• •• 

• • • 

• • • 



dP • 
OP h1 

(^ ted 



1 
1 


1 

1 
1 


1 
1 


1 
1 


04 


< 


Pd 


r4 



42 



For each line of braille text, there follows a count of the 
number of units in each of the following categories: <. 
ALPH alphabetic letters 
NONA non-alphabetic letters (numbers, 
decimal point) '^! t ^t> ■ ; 

OTHR non-alphabetic letters (asterisk, 

apostrophe ) 
GRAM phonograun, morphogram, logogram 
PUNC punctuation (modulation) 
RGST register (modulation) 
Note: These present categories closely approach the inventory 
discerned in Hamp & Caton, 1984. As our work 
progresses we intend to expand these categorial 
discriminations by basing the ongoing analysis on ever 
more refined observations . 
Detailed definitions of the categories can be found in the 
teachers edition of each level of Patterns: The Primary Braille 
Reading Program. The braille units in a line are listed 

individually by name and number of occurrences. See figure 1 for 
an illustrative reproduction of the above combined data. The 
program also prints a count of braille units individually and by 
category for each seimple and for the complete file; these data will 
be discussed below. 

Now that we have our sample corpus in an appropriate state for 
computer processing in both braille and print versions, and already 
with some facilitating analytic categorization correlated in 



43 



accessible form, it is immediately apparent that a great wealth of 
information lies amply awaiting imaginative consultation or else 
can be readily developed with the investment of only a modest 
cimount of further data insertion and analysis. Indeed, we already 
have at hand a shopping list of embarrassing proportions. Along 
these lines, we will welcome inquiries so far as we are in a 
position to satisfy them. 

As a minuscule sample of the information already in hand we 
offer here (Figure 2) a tabulation of the totals of braille units 
occurring in the entire set of 25 texts drawn from the Brown 
Corpus . 






44 



ir»'^oinmCT^oii~"<-i'*"<3'>-»r>-ino^'-''*i^'^'*<*)nmoo 
noo.-ii<-i<Ni rnicNfN miir* iin^r-i 

r- I VOll I CNII-HII II I 

I III I I I I I II I I I I I I I I I 



•-) EH rn tn 

iJ W O 

w o 



>^«i:EHCQtQOMUaiZ»^><00:! 



ca td m H 

Oi (0 Eh t^ 



ina>o\(NOOvo^o\f*)inr»iHrsmoooooooocN'-iin 
r-.Ot-<t-tinr~iHmtHfni-Hr--rom<-tnin<Nrot~-n<Ninr~ 
usncM r^t-t CM r-<-H co iH mooiH'V 



vo n CN r^ 

CO o I (N 

rH I I 



I I 



I I 



I 



I I I 



pLi P4 Q ^ Ck EH ^ 



Trr->vo^mr^mcotn^rsooor-4(N^vomiHmr-r-oocD 
minnmvx)<-ioi-<r» incNrotH vooor-itHvoin^n 

oonin 'a'l-i^ii i cN>-HrH 

•H r~- II I I I I I I I I I I I I I 

I I I I I I I I I I I I I I I I I I I I I I 



oofH ^EHonoa eh h 
Q u m << u b 



O D3 



u EH Q m D >i q 
g (^ 01 u (x; oi K 

ph eh > s 



r~ 
f-t 

EH 

O 

&H 



ooooo>oocNvoiH<Nr-nnr~cNO^oinr~inCT<(T(mna>^ 

OOOOtNrMONintNiHfNCN'HtnCT^'-linCNO incNncNtN 

<N(NC>j>-i rn rHcsn r-i »h 

O) I III II I II I I I I I 

I I I I I I I I I I I I I I I I I I I I I I 

tjuo:ifaS(aEH(^(o>4Scn:3ai>4iQEHKiu)cQ«« 

UMEHt<^ pQnQ> mSEHO OEhO eh 



o do 



0) 

& 

•H 
P4 






EH 

Vi 
ID 



1^ 



u 



Oi 



vo^oo^vor-inr»^cnrnoovor<rors<-4inrHooiri(N 
f-imoor-cNO\fnnTi« cn ^vocNniHmo\ ooocso^ 
oirtr-Ht-too roi-Hooir-ir-i iH^rmivofN cn 

(N^ oil rHI I III t-< lr-4 I 

II III I I I I I I I I II III 



1x4 P4 t3 a» 



tfS>i«(oaa»:]tfOhEHUpM^OEitosm 

(^P^»q;^|x|OUUOS(0KOO>0]HOEH 



« EH D i2 



o\vo«onioi-i.-itn(Noni-HcNr»r-ivor-<o>ino\oot-»inr*ro 
^^o\oon cn^rorHo* invonrOrH^r^i-t^oovooo 
r^voi^(Nlr-tH iini-tcNfOfM i 

(Nt^CSl I I IIICNI lllllll 

•H I I I I I I I I I I I I I I I I t I I I I 

HO>HCOUOiSEH>«2U>0::t^S»HEHEHEH>EHtOPi;f^b 

Hqcocoosn o 1^ HcqsopiiP^mzHX 



'^ 


ooocot^Ti<^OrHo\'*oovoo\oir>r-tnonn{N^nin«) 


CO 


vor^vooooo i-incN<-i vo^hcn r-irrn ,H(nr> fnm 


n 


r^orr cnioj ^ I cmi ^hi i-i.-hi 


n 


roin 1 1 iMiiii III II III 


>* 


1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 



OZ»<r-»<;OiSOQ>:iEHQtaS 

EHU)e>)<ZHZOOO.. 

«i; 1^ »q u H ii« n 




(o>»;Gq(t:S(opi 

(OKmSHEHOX 

5 EH o n m 

O Pi (0 EH g 



03 

EH 

o 



<y>rvji-HvnrniHvooiHr»o%vo«niniH 
r>-ooo'<)'vocN'H>-H r«o\voiHin 
CT\ m vo "H n i-< I iH 

■^ ■* <-H I II III II 

I I I I I I I I I I I I 

ua^voPKCxEHSum 

h) OQ H fa U U 



onoocN«Hn<-<'a"CNvo 

iH O <H O I-* rH r» 

fH I I iH rH CN I 

II I II I 

I I I I I I I I I I 



XM«W(0QHEHWO>0{mO 
.. _„,<<UI«4«EHKW2EH(05:OUa 

f^qn«t]j2(Quoz MH025O nEnnx 






o 



"tpcNr-tni-Hmr-vocNOCNvoovor-r-^nt^oor-THCTivotN 
^voo\r~ voir>incN^^r«t-HrocNr~-'a' ih voorHinvo 

(N<nO>rH|r)<-H CS 0\«HCN^ liHIi-HrH rH 

CNl^»-H I III ItH II I II 

I I I I I I I I I I I I I I I I I I I 



cQH:q>m(0C)i^faSfa>H 
Pj n A« nj 
S O Eh 
5 h^ H 



Q Z •« 



fJ H CQ CJ U Oi t> 



I 



EH^JCJQtCStfOPiiD 

Ht^UZOOHEHHO 

z 5 H m m X 
O (0 EH g 



vo 
o 
n 
(*• 



04 



oonnofsrorHoncNCOi-Hvooki-HvocoinnrHovoiHCNO 
t--r-'^o\0'^r»rH ino\i-HiHvo(^ooinf*)inrMtnvo'^rnoo 

~ '~ " ^tnCN COfH^rH 

I 'a" rH 

I I I 



r»- r- '^ o\ o ^ r- 

o\ (N a> in n 

O rH n I I 

rH I I I I 



rt; « D '* 



Eh 04 Oi 

04 CJ u 

X n << 

U Q 



I I 



I I 



I I I 



I 



I 



U Eh Z H 05 
H D O H O 

0) n u fa 



EH Z Q M o to q 

O Eh 03 n EH ? S 

. D (0 (0 Eh 
ZOO 



45 



Even a casual glance at the totals for the 25 texts reveals a 
dramatic fact. It is known that the frequencies for whole English 
words in the Brown Corpus show at the lower end a long trail off 
with many single occurrences. Out of 37851 dictionary words (i.e., 
forms which would appear as headwords in a dictionary) which make 
up the Brown Corpus 2124 account for 80% of the Corpus text, while 
as many as 22000, or 58% of the list, occur but once each. Note 
that the pattern for least frequent braille units is entirely 
different. 

We propose to continue in the near future with studies of this 
corpus sample more penetrating and less obvious, or superficial, in 
character and more promising for the solution of problems that 
previously a sharp eye and pencil could scarcely touch. 



46 



References 

Akmajian, Adrian, Demers, Richard A. , & Harnish, Robert M. (1979). 
Linguistics: An introduction to language and communication. 
Cambridge, MA and London: MIT Press, Pp. xvii + 357. [The 
theoretical basis is predominantly that attributed to Chomsky 
in the later 1970s; substantially revised as A.A. , R.A.D., Ann 
K. Farmer, and R.M.H., 3rd edition 1990, Pp. xiv + 508.] 

Caton, H., Pester, E, & Bradley, E. J. (1980). Patterns: The 
primary braille reading program. American Printing House for 
the Blind, Louisville, KY 

Crane, L. Ben, Yeager, Edward, & Whitman, Randal L. (1981). An 
introduction to linguistics. Boston and Toronto: Little, 
Brown, Pp. xiv + 280. [Accessible yet informative, fairly 
eclectic in theory; but mainly of mid-7 Os vintage] 

Davies, Anna Morpurgo. (1991). Historical Linguistics, 
International Encyclopedia of Linguistics , Oxford 1991. 

Finegan, Edward, & Besnier, Niko. (1989). Language: Its 
structure and use. San Diego, New York, etc.: Harcourt, 
Brace, Jovanovich, Pp. xiv + 546. [Broadly informed] 

Hamp, E. P., & Caton, H. (1984). A fresh look at the sign system 
of the braille code. Journal of Visual Impairment & 
Blindness, 78, 210-214. 

KuCera, H., & Francis, W. N. (1967). Computational analysis of 
present-day American English, (generally Icnown as the "Brown 
Corpus"), With a foreword by W. F. Twaddell, a study by Mary 
Lois Marclcworth and Laura M. Bell, and an analytical essay by 



47 



John B. Carroll. Providence, Brown University Press. 

Lyons, John. (1968). Introduction to theoretical linguistics, 
Ccunbridge University Press. Pp. x + 519. [durable, original, 
literate] 

O'Grady, William, & Dobrovolsky, Michael, (Ed.) Aronoff, Mark. 
(1989). Contemporary linguistics: An introduction. New 
York: St. Martin's Press. 

Pedersen, Holger. (1924). Sprogvidenskaben i det nittenae 
aarhundrede: metoder og resultater. Kjobenhavn, Denmark: 
Glydendalski Bognandel Nordisk Forlag; (1931) trans. Spargo, 
J. : Linguistic Science in the nineteenth century. Cambridge, 
MA: Harvard; reprinted with changed title: (1962). The 
Discovery of language . Bloomington: Indiana University Press 
(Midland Book) . [Authoritative classic by a magisterial 
participant; see p. 289 of trans., 266 of original for picture 
of F. de Saussure; and p. 63 of trans, for picture of and 
reference to work of Pedersen who did not include these in his 
original] 

Senner, Wayne. (Ed.) (1989). The origins of writing. Lincoln 
and London: University of Nebraska. Pp. viii + 245. 
[Authoritative presentations of eleven writing systems by 
specialists; eminently readable] 



48 



An Analysis of Braille Word Lengths 

Eric Hamp 

John Si ems 

Hilda Caton 

American Printing House for the Blind 

Louisville, KY 

1993 



49 



Abstract 

Lengths of braille words were studied as part of the linguistic 
analysis of Grade 2 Braille. Base data consisted of Brown Corpus 
selections, which have been described in the preceding study. In 
recognition of the known complexity yet finite nature of the notion 
"word" some of the text items were excluded in order to yield an 
indisputable body of "normal" units for the study. Excluded, or 
"Selected," text segments comprise: 1. items of arbitrary length 
and structure such as numerals and acronyms; 2. aberrant items, 
including hyphenated complexes and proper names, which raise 
problems of criteria; and 3. debatable items such as 
abbreviations, print contractions, foreign words, titles, and day 
and month names. There remained 45408 running text words from 
which punctuation and composition signs were removed as fortuitous 
or irrelevant. Each word was assigned a length consisting of three 
numbers in the format C,N,S, giving the counts of print characters, 
braille units, and braille shapes. For example, the length of 
tension is 7,3,4. Various tabulations were made of these lengths. 
While there is a preponderance of short words in both print and 
braille, this feature is more marked in braille. More than one- 
third of the words in braille text consist of only one shape. 
While 43 percent of print words have three or fewer letters, in 
braille 63 percent have three or fewer units. Certain length 
combinations occur very freguently. More than 19 percent of all 
words have the length 3,1,1. Five lengths (3,1,1 2,1,1 2,2,2 
4,1,1 and 4,4,4) account for 40 percent of the words. Words which 



50 



have no braille contractions, i. e., where the counts of C, N, and 
S are equal, constitute ca. one-fourth of the text. In more than 
one-half of the word lengths, one finds one or two fewer shapes in 
braille than characters in print. Tables of interrelation among 
the counts of print characters, braille units, and braille shapes 
confirm our intuition that while there is general correspondence 
between the lengths of words in print and braille, deviation arises 
noticeably out of the existence in braille of units (contractions) 
which may stand for anywhere from two to ten letters . On the other 
hand, between braille unit counts and braille shape counts a close 
parallel prevails. Word length information is important for 
braille education and for an understanding of the characteristics 
of the braille communication medium and its genres. 



51 



An Analysis of Braille Word Lengths 

A study of the lengths of words found in running American 
English braille text has been conducted as part of the project "A 
Linguistic Analysis of Grade 2 Braille." 

Data for the study comprised a braille version of selections 
from the Brown Corpus. This material, used also for other analyses 
in the project, consists of 25 sample texts of equal length 
representative of the range of subjects and styles in the Brown 
Corpus literature; cf. no. II of this collection. 

The notion "word" may seem at first blush to be rather simple 
and obvious: perhaps what occupies the line between two spaces on 
a page. What then do we do with sequences such as pre- and 
postmodernism or brother- and sister-in-law? If we isolate pre- we 
not only end up with a "word" that offends our sense of notional 
integrity, but we embarrass our analytic criteria by extracting a 
"word" without a "base", without its core. With brother-, even 
worse, we allege an erroneous kin relation. Let us then reverse 
our stance and register the full part which was elided; but this is 
surely false, since the elision was deliberate and its purpose was 
to avoid unwanted repetition. Besides, such a procedure tampers 
with the text; we then have no easy or objective way of forbidding 
the gratuitous supply of all sorts of addenda which a reader might 
feel to be explicatory or even as clarifying and more felicitous 
for the style. Our job is to count what is there, to have explicit 
rules for recognizing what is there and for keeping track of what 



52 



we do not count. 

There are other ways in which the search for a satisfactory 
definition of the notion "word" encounters difficulty. Are both 
story and stories instances of the same word: are sleep and slept, 
or feed and fed, the same word; or go and went? If we agree that 
story and storey are different words, what about dialog and 
dialogue? Is {to be) heading the same word as (to) head, or head 
(home) the same as head {the class)? Is heading {the class) an 
instance of (a newspaper) heading? Is it useful to say that 
headstrong CONTAINS the word head, or that timeliness CONTAINS time 
and timely? That is, can words be inside words? If so, what about 
heady or hearty? 

There are even more subtle problems in deciding what a "word" 
is and what we should regard as its extent. It is easy to see that 
the verb forms leans, leaning, and leaned have a part in common. 
It is a bit less clear-cut, if we take pronunciation into account, 
to claim that means, meaning, and meant (all taken as forms of the 
verb) share just the same corresponding part. We see also that we 
can identify a common part in story and stories successfully if we 
adopt a simple rule that says that we must write i as y when it 
comes before a space. Let us consider now the simple set four, 
fourteen, and forty. If we can solve the problem mentioned above 
suggesting that words may contain words, then we find no 
predicament in saying that fourteen actually CONTAINS four (4+10 
with ten in a special shape, the vowel with two e's), and that 
forty has 4 and 10 in special shapes when the meaning of the 



53 



combination is ^multiply'. In other words, we will say that these 
derivative "words" are formed by simple adjoining. By stretching 
our criteria we can even account in a similar fashion for some of 
our verb forms: We might say of the pair leans and lean that the 
former has an -s for the third person {she, he or It) and that the 
latter then has lean as the skeleton or essence of 'leaning' or 
.^tilting' to convey the activity or state for all unspecified 
persons. In this way we have arrived at the common part of leans, 
leaning, leaned, and lean — a kind of abstract entity LEAN which 
denotes a certain verbal function stripped of its inflections. 

Yet things do not always work out so conveniently. If we 
consider another simple pair, peach and peaches, we cannot extract 
peach from peach-es in the same way without doing violence to the 
meaning. To put it simply, a peach is singular, but peach-es is 
not singular and plural. This is not to say that we cannot contrive 
to solve this riddle somehow; and if we remember a peach tree 
{never a peaches tree)', and notice with Peachtree Street in Atlanta 
how we waver with compounds) we are led to the essence, or abstract 
entity, of the noun PEACH. But we are still left with the task of 
stating just how peach denotes both PEACH and singular. An 
analogous problem presents itself in the case of meat, a mass noun 
{some meat, not a meat), and meats 'kinds of meat'. This is, of 
course, the problem that underlies a portion of the construction of 
a dictionary entry. 

We see then that a "word" is a complex notion. It has both a 
shape (a sound, or a written or brailled, configuration) and a 



54 



meaning or signified value. There are at least two kinds of 
annoyance that arise for us as we try to identify, delimit, 
measure, and count words in a written or brailled text. There is 
the question of what happens to be found standing between spaces 
and how deviant these spaces may be; and there is the problem of 
defining identity and function among partials within such non- 
deviant spans . 

There exists a very large serious literature touching on the 
notion "word" , too large for us to pretend to survey or even list 
here. Perhaps as a seeming irony, there is no entry word in the 
10-volume Encyclopedia of Language and Linguistics (1994; volume 9, 
page 4989); it seems that the very complexity of the notion led the 
editors to an immediate partition of the subject, or to eliminate 
it as a proper rubric. Thus we find the entry word, phonological 
(ibid. 1994: 5007-9), where it is stated that orthographic spaces 
beg the question, that this unit is the smallest piece which can be 
pronounced alone with a stable meaning, and that "small words" 
(e.g. was or for), with their varying pronunciation, create a 
problem. In short, all the discussion here dwells on the spoken 
language and dismisses the graphic as problematic. Otherwise, we 
find in this major encyclopedia only word recognition and lexical 
access (5009-14), word- format ion: compounding (5021-6; cf . Katamba 
1993: 297), word-formation: neo-classical combinations (5026-8; 
a puzzling category), and word- format ion: shortening (5029-31); a 
surprisingly short inventory. Yet in the computer-produced subject 
index to this Encyclopedia 1/94 or ca. 1%, of the listing of all 



55 



the non-obvious questions regarding language is captioned word or 
a phrase including the expression word. On the other hand, much of 
interest to the notion "word" is to be found in the article clitics 
(vol. 2, pp. 571-6). As a complex technical example of an attempt 
to arrive at an exact delineation of a phonological word using the 
criteria of clitic function, consider the following preliminary 
definitional footnote of Spotts 1953: 

We have chosen to consider the intonational unit 
described by Miss Pike as a phonological word, defined 
among other characteristics (for which see her paper) by 
intonemic placement, since each intoneme (or intoneme 
sequence) ends a word. Phonological words consist 
morphologically of an optional proclitic followed by a 
simple or compound stem which in turn is sometimes 
followed by an infix and/or an enclitic. A simple stem 
is composed of a root followed by a stem formative. A 
compound stem is composed of two simple stems in sequence 
or an abbreviation of these. 

Instead of trying to survey here the vast literature on the 
notion "word", when, after all, that is not itself the primary aim 
of our present study, let us sample some informed and serious 
general statements on this accepted, clearly useful, intuitively 
grasped notion that seems to apply in practically all human 
languages that have been attentively investigated. For some 
concise and clear remarks in simple terms we may first note from 



56 



Crystal's (1987:91) entry words: 

Words sit uneasily at the boundary between 
morphology and syntax. In some languages 
— "isolating' languages, such as Vietnamese — 
they are plainly low-level units, with little 
or no internal structure. In others — 
"polysynthetic' languages such as Eskimo— 
word-like units are highly complex forms, 
eguivalent to whole sentences .... - 

Because a literate society exposes its 
members to these units from early childhood, 

we all know where to put the spaces apart 

from a small number of problems, mainly to do 
with hyphenation. ... 

It is more difficult to decide what words 
are in the stream of speech, especially in a ^ 
language that has never been written down . • 
And we may excerpt from Crystal's (1987:104) section on "semantic 
structure" : 

People readily talk about the "meaning of 
words'. However, if we wish to enquire precisely i 
into semantic matters, this term will not do.... 
1. The terra word is used in ways that obscure the 
study of meaning. The forms walk, walks, walking, 



57 



and walked could all be called 'different words'; 
yet from a semantic point of view, they are all 
variants of the same underlying unit, 'walk' . If 
the variants are referred to as 'words' , though, 
what should the underlying unit be called?... 

2. The term word is useless for the study of 
idioms, which are also units of meaning.... 

3 . The term word has in any case been appropriated 
for use... in the field of grammar, where it does 
sterling service at the junction between syntax and 
morphology. ... 

For such reasons, most linguists prefer to talk 
about the basic units of semantic analysis with 
fresh terminology, and both lexeme and lexical item 
are in common use. ... It is lexemes that are 
usually listed as headwords in a dictionary. 

A more sophisticated statement concerning the perplexing and 
complex nature of the criteria for discerning these units is to be 
found in the article Words by E. M. Uhlenbeck in Bright (1992, vol, 
4: 246-8), and references therein. We excerpt here some leading 
parts of that excellent article, which should be consulted in its 
full text and followed up through its informed references. 

For the sentence as well as for the word, 
many definitions have been proposed; but so 
far none has gained general acceptance. . . . 



58 



This lack of consensus among linguists stands 
in sharp contrast to the general agreement of 
native speakers everywhere, who seem convinced 
that they have words at their disposal for 
daily use in actual speech. 

Four basic issues are involved in the 
problem of defining the word: 

1. Is the word a universal? That is, 
... is it part of every language? . . . 

2 . What kind of unit is a word? Are 
words grammatical units . . . ? Or are they 
primarily SIGNS, i.e. units of form and 
meaning. . . ? 

3. How are words related to sentences? 
Is every sentence. . . analyzable into a 
sequence of words...? 

4. What is the position of the word in 
language structure? Does it occupy, as was 
traditionally assumed, a central position in 
grammar . . . ? .^ ^ 

The answers given to all these questions 
largely depend on the theoretical position 
adopted. On the European continent, a 
majority of linguists tended to accept 
Saussure's conclusion that... (Saussure 
1916:159) seems to agree with what laymen 



59 



generally feel to be true. The Prague School 
as well as Dutch structuralists... built on 
this view; ... 

In Great Britain, however, . . .the word 
was viewed as a grammatical unit .... Other 
linguists seemed to have some reservations 
about the universality of the concept:... 

In American linguistics, the Neo- 
Bloomfieldians— influential between 1940 and 
19 60— generally considered not the word but the 
morpheme as the smallest and basic grammatical 
unit.... The question of universality was 
rarely broached before 1970. . . . 

Chomsky 1970 marks the beginning of a new 
period in which the lexicon and morphology, 
fields previously neglected in generative 
grammar, were recognized as important domains 
of linguistic research. . . This led to a 
renewed interest in the word, its role in 
grammar, and its semantic nature. This has 
resulted in psycholinguistic studies ... ,and 
studies within a generative framework—. . . 

Since the word may be the locus of 
morphonological, morphological, and syntactic 
regularities, it seems unlikely that one 
single criterion will suffice to distinguish 



60 



words from all other linguistic entities. 

The great Danish linguist Hjelmslev (1970, i.e. 1963 from ca, 
1943:32) provides us with a particularly pure Saussurean statement: 

Every language appears to us first of all 
as a system of signs, that is to say, a system 
of expression units that have content, or 
meaning, attached to them. Words are signs of 
this sort. But parts of words can also be 
signs :-s in English is a sign of the genitive 
(Jack'-s father) and a sign of the third 
person singular present (he write-s) . [the 
example in the Danish original is better and « 
clearer in the latter instance, since the -s 
is the Danish passive] A word like in-act-iv- 
ate-s [the Danish example more closely 
resembles un-like-ly , with three parts] is a 
sign consisting of five different smaller 
signs. A sign may consist of one expression 
element with one content element attached to 
it, like the English sign -s in Jack's father, 
which consists of the expression element s 
with its attached content element 
"genitive' ; or it may be formed-both on the ; 
expression side and on the content side— by the 
combination of two or more elements, like the 



61 



Latin sign — arum in bon-arum mulierum ^of the 
good women', which consists of four expression 
elements — a, r, u, and m — and three content 
elements—" genitive ' , " plural ' , and " feminine ' . 

And he follows this on page 91 with the ensuing 
characteristically exact and laconic observations : 

The only linguistic typology to achieve a 
place in classical linguistics was a 
classification according to linguistic usage. 
The central point of interest was the 
structure of signs, especially of words. 
Vlords are permutable signs, signs that can 
exchange places within a linguistic chain: 
softly answered consists of two words, because 
one can also say answered softly; soft-ly and 
answer-ed each consist of two signs, but these 
signs cannot be put in another order. 
Permutable signs attracted an extraordinary 
amount of attention from classical 
linguistics, beginning with antiquity, since 
it was thought, in connexion with Aristotelian 
conceptual logic, that each such sign stood 
for one concept. 

A more natural English example might be slowly nodded and 



62 



nodded slowly . We cannot replicate Hjelmslev's elegant Danish 
example of "the boy runs', where the Danish article attaches to the 
end of the noun. 

The seeming conflict in criteria invoked for the formal 
recognition of so widespread a notion as that of word is evident in 
the best theoretical linguistic writing of all periods and of most 
schools of thought. It should therefore not surprise us that a 
recent and well received handbook (Akmajian, Demers, Farmer, and 
Harnish 1990:11-13) tells us that words are encoded through 
features of phonology, of morphology, of syntax, of semantics, and 
of pragmatics — ^indeed of all the subfields of linguistics discerned 
and dealt with in that book. 

In view of the difficulties encountered in arriving at a 
satisfactory definition of a "word, " some text items were 
eliminated in order to assemble a body of data which might be 
generally agreed upon as consisting of "words." 

Moreover, for this study of word length it is of central 
importance that the material analyzed be composed of strongly 
finite segments . It is a part and an implication of the structure 
of language that its elements, such as words, are formed by rules 
that may indeed be iterative but that they are formed, or stored 
(e.g. in the dictionary), in lengths that are not unlimited; these 
lengths are in fact governed by the rules of grammar that 
specifically characterize each language. It is therefore of 
negligible interest to our enquiry if a written formation can be 
found that in principle may be of unlimited or inordinate length. 



63 



For this study, the corpus of text items, in relation to 
spacing, has been dichotomized into "selected" segments and "normal 
words." The "selected" items have been excluded from the study; 
the "normal words" are viewed as our first priority for study, 
especially since they, after all, must be the prime object of 
learning and teaching for braille instruction and since they will 
potentially occur in almost any English text, i.e. they are as a 
set maximally context-free. They most centrally characterize 
English. 

As in previous processing of the Brown Corpus samples, 
apparent errors are not corrected. The KuCera analysis seems to 
use the same approach of letting errors stand. In cases in which 
the classification of an item is questionable it seems better to 
lean toward "selecting" the item (that is, excluding it from the 
study) so that as far as possible the remaining text can be viewed 
as a defensible base of normal words. 

The "selected" items which have been excluded include' the 
classes below. 

We intend at a later date, when time permits, to study the 
numerical effects of these exclusions, especially by category of 
"selected" items. It is already apparent to us, as we reflect 
since initiating our task of selection and counting, that our 
totals at that time will change very slightly as we revise some 
individual judgments. But these changes will be minimal, and at 
present would necessitate an entire wasteful recalculation. 



64 



1. Those with arbitrary length and constituency. 

Niunerals: By definition these may be undefinably large. They 
involve signs which are discretely different from linguistic 
word markers. Moreover, it is difficult to predict the number 
of English words which will surface; e.g. a hundred and five 
or 073e hundred five. We refer here to Arabic, Roman and 
letter-number combinations . 

Acronyms: One cannot always say how many (small grammatical) 
words are represented. Sometimes the acronym itself is meant 
as a word; often not. A very chancy class. 

Mathematical formulas: These, represented in the Brown Corpus 
by the symbol "**f," are arbitrary with respect to our study 
since they obey mathematical but not linguistic syntax. We 
also cannot predict precisely how they will be vocalized in 
language . 

Ellipses: The amount of text so treated is not always easy to 
specify or depends upon parsing. We cannot classify zero 
here. Perhaps suspension points can be classed with 
punctuation. .- .> 



65 



Systematically aberrant. 

Printer ornaunents: These are not language, in a strict sense. 

Greek letters: Representation of these letters in Roman 
typography requires the use of arbitrary multiple-character 
symbols. Transliteration may not be consistent. 

A 

Hyphenated words: fade-in, half -intensity . Compounds which 
are partially hyphenated and partially spaced are treated as 
hyphenated. Examples: often-blood thirsty, mid-twentieth 
century. English also has no upper limit: quasi — pay as you 
go. Where does the hyphen belong? Words may be written with 
hyphens between letters, such as w-i-d-e. Like pig-Latin 
these are oddly derived from "normal words." 

Proper names: Names of persons; headings and titles of 
articles, plays, books, other publications; names of 
geographical places and locations; names of governmental units 
and other organizations; names of specialized equipment. 
Names in English and modern styles (e.g., literary titles) 
have few limits. Names refer, but do not mean; this can offer 
some problems of judgment. When is a place a name? There is 
also an old philosophical problem of uniqueness {the sun — is 
that a name?). Geographic names (and ethnica such as the 
!Kung) can be "foreign" (see below). Names of organizations 



66 



can be wordy and expansible. How many words are there in a 
full personal name? 

Letters representing mathematical variables: These items may 
be somewhat like proper names. In the phrases "the point p" 
and "the horse Secretariat" are not the terms "p" and 
"Secretariat" linguistically analogous? 

3. Debatable items. 

Abbreviations: Although some of these embrace multiples of 
words (i.e., e.g.), or are of ambiguous rendering (e.g. = e 
g or for example; ca. = circa or about or approximately or 
around) , others have taken on a life of their own as words 
{etc. as an adverb etcetera, with a variant ex(c)et(e)ra among 
others) . 

Print contractions: In cases such as don't or and/or it is 
not perfectly clear whether we are to assign the count before 
or after contraction. A form such as whaddya, which tends to 
belong to a style reflecting colloquial spoken varieties, 
really reflects a separate class of "folk phonetic" spellings. 



67 



Foreign words: In one sense these could be called 
systematically aberrant and dismissed as such; it could be 
claimed that their inclusion in our count would imply an exact 
knowledge of all the world's languages. Yet there are 
specimens that have surely been incorporated in our language 
(wadi, ladino, samurai, maharaja, kanaka^, aa^, cosa nostra^, 
Zeitgeist , hajj, mukluk, kachina, milpa) . The indecision that 
our tradition systematically imposes on us is considerably 
alleviated for German speakers, who call naturalized 
borrowings (of any age) Lehnworter "loan-words," but adoptions 
which still carry a foreign flavor Fremdworter "foreign- 
words . " In questionable cases the braille rule of identifying 
as foreign those words which are in a different type face is 
applied. 

Titles: These form a troublesome class. In combination with 
names they might be considered as components of proper names . 
Yet alone they seem to function as "normal" appellatives. 

Days of the week and month names: Though these can well, on 
rules of capitalization, be regarded as proper names, they 
also, as grammatical classes with finite membership, can be 
taken as a subtype of "normal words." 



68 



We may therefore note residually the classes of items not 
"selected," i.e. text material treated as part of the group of 
"normal words": 
Words in vocabularies of specialized subjects: : lats , reps. 

Dialectal representations: git, 'scuse. 

Historic contractions: The combination of cabin and hut contained 
in the word cahoots has somewhat receded into the past. 

Spaced Compounds: While for this study, the foregoing items, in 
general, can be identified by simple inspection, a somewhat 
different type of problem, for which proper identification may 
require a considerable amount of judgment, is presented by the 
spaced compounds . These items are included in the count and no 
doubt constitute an exception to the principle of "general 
agreement" regarding the content of the data upon which the study 
is based. In such compound terms as grand jury and social 
contract, the elements, grand, jury, social, and contract, are 
treated as individual words even though our native linguistic sense 
might lead us to regard the whole span as a single semantic word. 

Since the focus of the study was primarily upon the words 
found in the text rather than upon the discourse structure of the 
text, capitalization, italicizing, and punctuation markings were 
ignored . 



69 



< ' The text samples, after modification, contained a total of 
45408 words. In this total and in the following tables, a "word" 
corresponds to what is called a "token" in the Brown Corpus 
introductory description. That is, for multiple occurrences of the 
same lexical word, each occurrence of the word in the text is 
registered in the count. 

In order to make the tabulations, each word was assigned a 
length expressed as a three-number combination: C,N,S. The first 
number is the count of the print characters in the word. The 
second is the count of braille units in the braille translation of 
the word. The third number is the count of braille shapes. For 
example, for the word father C,N,S has a value of 6,1,2. It 
contains six characters in print; in braille it consists of one 
braille unit and has two braille shapes. The length of the word 
tension can be expressed as 7,3,4. It has seven letters, three 
braille units, and in braille is written using four shapes 
occupying four cells. 

The first Table gives the print character lengths, braille 
unit lengths, and braille shape lengths of the words found in the 
text. Specifications of lengths are contained in column L. Column 
C shows the number of text words having print character lengths 
equal to the specified lengths in L. Column N shows the number of 
words which, when translated into braille, have lengths in braille 
units equal to L. In column S are found the number of words having 
lengths in braille shapes equal to the L lengths. The columns to 
the right of C, N, and S give the percentages of words having each 



70 



length and the cumulative percentages of words having lengths equal 
to or less than the specified lengths. 

Because of the many contractions (phonograms, morphograms, and 
logograms) which are a major feature of the braille code, braille 
words tend to have fewer characters than corresponding words in 
print. The analysis gives some details in regard to specific 
effects of this compression. ^j 

While the majority of words in both print and braille are 
relatively short, braille has a greater concentration of short 
words. As shown in Table 1, the most frequently occurring word 
length in print is a length of three characters. In braille the 
most frequent length both in terms of units and shapes is a length 
of one. In braille more than one-half of the words have two or 
fewer units or shapes . In print a maximum character length of two 
accounts for one about one-sixth of the words . In the Table there 
appear to be several points at which braille word lengths (N and S) 
correspond in frequency to print character lengths (C) greater by 
two. Approximately 90 percent of the words have six or fewer units 
and shapes but up to eight characters. Similarly, the percent of 
words which contain ten or fewer braille units and shapes is about 
the same as the percent of words made up of twelve or fewer print 
characters . 

We may paraphrase the foregoing observations in somewhat 
different quantitative terms: Over 1/4 of all inkprint words 
comprise 3 letters; there is then a bell-shaped decrement over 
nearly 20 letters of length. Braille shows a strong contrast in 



71 



this respect. Over 2/5 of all words, or 4 words in 9, consist of 
but one braille unit. For greater lengths the decrement is 
stepwise: to 5, then to 1 , to 9, then 10, and to 14 in exiguous 
numbers. Thus we see that braille units are 1/2 again (i.e. a 
ratio of 43.5 to 2 6.3) as economical as inkprint characters in the 
most frequently occurring length. In other words, the compression 
has been introduced in a highly useful set of items, and any 
improvement in the code in this regard would presumably have to 
consider items one by one in terms of text occurrence. 

With the exception of lengths of one, shape counts come close 
to braille unit counts; except for one interesting inversion at 
length 5, the shape count is always slightly higher than the unit 
count. Now we may remove the anomalies of lengths one and two 
substantially by summing them; this clearly results from the 
analytic fact that braille units often consist of two shapes. We 
may draw from this an important consequence of braille shape 
design: It appears that braille units derive their diversity 
(recognition characteristics) without exploiting undue shape 
complexity. To attain this diversity (=optimization of word 
length) otherwise would entail packing more information in single 
braille shapes. 



72 



Table 1 



cum. 

per- per- 

L C cent cent N 

1 1355 2.984 2.984 

2 6152 13.548 16.532 

3 11944 26.304 42.836 

4 7379 16.250 59.086 

5 5277 11.621 70.707 4406 

6 3307 7.283 77.990 3063 

7 3252 7.162 85.152 2126 

8 2525 5.561 90.713 1226 

9 1734 3.819 94.532 609 

10 1130 2.489 97.021 307 

11 693 1.526 98.547 147 

12 371 .817 99.364 54 

13 170 .374 99.738 23 

14 82 .181 99.919 5 

15 29 .064 99.983 

16 5 .011 99.994 

17 3 .007 100.001 



cum. 
per- per- 
cent cent 

19758 43.512 43.512 

4766 10.496 54.008 

4066 8.954 62.962 

4852 10.685 73.647 

9.703 83.350 

6.746 90.096 

4.682 94.778 

2.700 97.478 

1.341 98.819 

.676 99.495 

.324 99.819 

.119 99.938 

.051 99.989 

.011 100.000 



cum. 

per- per- 

S cent cent 

16360 36.029 36.029 

7415 16.330 52.359 

4284 9.434 61.793 

4904 10.800 72.593 

4315 9.503 82.096 

3165 6.970 89.066 

2264 4.986 94.052 

1339 2.949 97.001 

743 1.636 98.637 

358 .788 99.425 

168 .370 99.795 

58 .128 99.923 

30 .066 99.989 

5 .011 100.000 



73 



Table 2 lists the most frequently occurring of the three- 
number combination lengths assigned to the words. Length values 
are listed which characterize one percent of all the words or more. 

Table 2 



3,1,1 


8764 


19.301 


2,1,1 


3459 


7.618 


2,2,2 


2693 


5.931 


4,1,1 


2240 


4.933 


4,4,4 


2123 


4.675 


3,3,3 


1507 


3.319 


4,3,3 


1498 


3.299 


5,4,4 


1435 


3.160 


5,5,5 


1421 


3.129 


1,1,1 


1355 


2.984 


6,5,5 


1259 


2.773 


5,1,2 


1144 


2.519 


3,2,2 


995 


2.191 


7,6,6 


921 


2.028 


4,1,2 


867 


1.909 


6,6,6 


781 


1.720 


7,5,5 


765 


1.685 


3,1,2 


678 


1.493 


7,7,7 


648 


1.427 


8,7,7 


613 


1.350 


8,6,6 


609 


1.341 


6,4,4 


573 


1.262 


4,2,2 


543 


1.196 


5,1,1 


470 


1.035 


5,3,3 


451 


.993 



Of the various three-number lengths assigned to the words in 
the text. Table 2 shows that the length 3,1,1 (three print 
characters, one braille unit, and one braille shape) is by far the 
most frequent. Words of this type occur more than two and one-half 
times as often as words of any other length. 

Certain patterns in the most frequently occurring lengths 
suggested further tabulations . Table 3 shows the frequencies of 
words whose lengths have a configuration of C = N = S. In Table 4 
there are counts of words represented by one unit and one shape in 

74 



braille so that N = S = 1 





Table 3 




1,1,1 


1355 


2.984 


2,2,2 


2693 


5.931 


3,3,3 


1507 


3.319 


4,4,4 


2123 


4.675 


5,5,5 


1421 


3.129 


6,6,6 


781 


1.720 


7,7,7 


648 


1.427 


8,8,8 


390 


.859 


9,9,9 


197 


.434 


10,10,10 


119 


.262 


11,11,11 


57 


.126 


12,12,12 


10 


.022 


13,13,13 


8 


.018 


14,14,14 


1 


.002 




11310 


24.908 



According to Table 3 about one-fourth of the words have an 
equal number of characters, units, and shapes. These lengths 
represent words which do not contain braille contractions, i.e., 
are composed of alphabetic letters. 

Table 4 



1,1,1 


1355 


2.984 


2,1,1 


3459 


7.618 


3,1,1 


8764 


19.301 


4,1,1 


2240 


4.933 


5,1,1 


470 


1.035 


6,1,1 


68 


.150 


9,1,1 


4 


.009 




16360 


36.030 



The most prevalent three-number lengths also indicate that 
many words have one or two fewer components in braille than in 
print. In Tables 5 and 6 counts of these words are listed. 
Lengths in Table 5 have the pattern N = S = C - 1 . Table 6 
displays the numbers of words having aN=S=C-2 length pattern. 



75 





Table 5 




2,1,1 


3459 


7.618 


3,2,2 


995 


2.191 


4,3,3 


1498 


3.299 


5,4,4 


1435 


3.160 


6,5,5 


1259 


2.773 


7,6,6 


921 


2.028 


8,7,7 


613 


1.350 


9,8,8 


358 


.788 


10,9,9 


184 


.405 


11,10,10 


102 


.225 


12,11,11 


55 


.121 


13,12,12 


22 


.048 


14,13,13 


9 


.020 


15,14,14 


3 


.007 




10913 


24.033 





Table 6 






3,1,1 


8764 


19 


.301 


4,2,2 


543 


1 


.196 


5,3,3 


451 




.993 


6,4,4 


573 


1 


.262 


7,5,5 


765 


1 


.685 


8,6,6 


609 


1 


.341 


9,7,7 


388 




.854 


10,8,8 


182 




.401 


11,9,9 


93 




.205 


12,10,10 


35 




.077 


13,11,11 


21 




.046 


14,12,12 


12 




.026 


15,13,13 


6 




.013 


16,14,14 


1 




.002 




12443 


27 


.402 



From the list of the most frequently occurring lengths (Table 
2) it is notable that many words are translated into one unit and 
one shape in braille. Table 4 shows that the total number of these 
words is 36 percent. Other prominent configurations in the list of 
most frequent word lengths are those in which the number of units 



76 



and shapes is one less or two less than the number of print 
characters . Tables 5 and 6 show that words having these length 
combinations make up more than one-half of the text. Tables 5 and 
6, of course, include some counts from Table 4. 

The last three tables present the interactions among the three 
factors: print character length, braille unit length, and braille 
shape length. In Table 7, the numbers to the left are character 
lengths and column headings at the top are unit lengths . 
Similarly, Table 8 shows the relationships between character 
lengths and braille shape lengths, and Table 9 shows the 
relationships between braille unit and shape lengths. From the 
three tables may be found the word count for any combination of two 
of the three length factors, C, N, and S. Each table is followed 
by diagonal totals which serve to summarize relationships among the 
factors . 



77 



Table 7 



78 



Table 7 



N 



10 11 12 13 14 



1 1355 

2 3459 2693 

3 9442 995 1507 

4 3107 651 1498 2123 

5 1662 235 524 1435 1421 

6 372 77 218 600 1259 781 

7 252 77 159 401 794 921 648 

8 60 18 54 154 580 656 613 390 

9 35 8 67 69 228 362 410 358 197 

10 14 10 28 30 89 223 243 190 184 119 

11 2 11 31 17 83 104 173 113 102 57 

12 5 14 22 69 65 87 44 55 10 

13 4 8 30 30 25 21 22 22 8 

14 4 2 7 20 3 14 8 14 9 1 

15 52 72463 

16 13 1 

17 2 1 



N 


= 


C 




11310 


24.908 


N 


= 


c - 


1 


10913 


24.033 


N 


= 


c - 


2 


13467 


29.658 


N 


= 


c - 


3 


5439 


11.978 


N 


= 


c - 


4 


2716 


5.981 


N 


= 


c - 


5 


855 


1.883 


N 


= 


c - 


6 


458 


1.009 


N 


= 


c - 


7 


156 


.344 


N 


= 


c - 


8 


65 


.143 


N 


= 


c - 


9 


29 


.064 



Table 8 



10 11 12 13 14 



1 1355 

2 3459 2693 

3 8764 1673 1507 

4 2240 1410 1606 2123 

5 470 1193 685 1508 1421 

6 68 273 191 708 1286 781 

7 136 223 375 920 950 648 

8 26 39 106 435 869 660 390 

9 4 11 21 59 152 395 515 380 197 

10 10 24 58 121 299 307 192 119 

11 2 1 38 32 95 154 192 122 57 

12 1 12 35 73 102 74 64 10 

13 4 1 6 29 36 34 29 23 8 

14 4 1 6 22 9 10 18 11 1 

15 5 2 8 1 10 3 

16 4 1 

17 2 1 



s 


= 


c 




11310 


24.908 


s 


= 


c - 


1 


11937 


26.288 


s 


= 


c - 


2 


14502 


31.937 


s 


= 


c - 


3 


5429 


11.956 


s 


= 


c - 


4 


1571 


3.460 


s 


= 


c - 


5 


480 


1.057 


s 


= 


c - 


6 


135 


.297 


s 


= 


c - 


7 


25 


.055 


s 


= 


c - 


8 


19 


.042 



Table 9 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 

N 

1 16360 3109 274 15 

2 4306 420 39 1 

3 3590 404 72 

4 4446 390 16 

5 3852 544 10 

6 2605 453 5 

7 1801 317 8 

8 1017 209 

9 ' ^ 526 83 

10 275 32 

11 136 11 

12 47 7 

13 23 

14 J 



S = N + 3 16 "^ .035 

S = N + 2 424 .934 

S = N + 1 5979 13.167 

S = N 38989 85.864 



Table 8 



79 



Tables 7 and 8 indicate that while there is a correspondence 
between word lengths in print and braille, for any one print 
character length a wide range of braille unit lengths or braille 
shape lengths may be found. For example: words having nine 
characters may be represented by anywhere from one to nine units in 
braille; words of eleven characters on a print page may occupy from 
three to eleven shapes in braille. 

There is rather close correspondence between braille units and 
braille shapes as in evident from Table 9. More than 85 percent of 
the braille words in the study have the same number of shapes as 
units. 

The tabulations resulting from this study provide a profile of 
word lengths in braille as compared with word lengths in print 
text. It is hoped that this information will be of interest and 
helpful for aspects of braille education. 

An ability to specify the range of length found in words in 
text constitutes for a language such as English an important 
characterizing feature of the nature and texture of discourse and 
written matter in that language. 



80 



Table 9 



81 



Notes 

'Our choice of the vague word "include" is deliberate. Anyone who 

sits down seriously to this job will find rapidly that the borders 

to judgment become perplexing, and the more one consults reflective 

philosophers and linguists the more the fuzzy and gray boundaries 

multiply. 

^Listed in the 1968 Random House Dictionary, College Edition, but 

not in Webster's Ninth New Collegiate Dictionary (1983). 



82 



References 
Aronoff, Mark. (1994). Morphology by Itself: Stems and 

Inflectional Classes. Cambridge, MA, London: The MIT Press. 
Asher, R. E. and Simpson, J.M.Y. eds . (1994). The Encyclopedia 

of Language and Linguistics . Oxford, etc. : Pergamon. 
Bright, William O. ed. (1992). International Encyclopaedia of 

Linguistics. Oxford: Oxford University Press. 
Crystal, David. (1987). The Cambridge Encyclopedia of Language. 

Cambridge: Cambridge University Press. 
Hjelmslev, Louis. (1970). Language: An Introduction. (Sproget, 

translated from the Danish by Francis J. Whitfield) Madison, 

Milwaukee, WI, and London: The University of Wisconsin Press. 

[the original published in 1963 in Kobenhavn, was actually 

written ca. 1943] 
Katamba, Francis. (1993). Morphology. New York: St. Martin's 

Press. 
Saussure, Ferdinand de. (see full reference, in the first 

study in this volume, pp. 00-00). 
Spotts, Hazel. (1953). "Vowel harmony and consonant sequences in 

Mazahua (Otomi')", International Journal of American 

Linguistics. 19: 253-258. 



83 



The Text Frequency and Incidence 
I of Certain Lower-Sign Sequences in American Braille 

' Eric Hamp 

John Siems 
[ Hilda Caton 

American Printing House for the Blind 

Louisville , KY 

1992 



84 



Abstract 

The so-called lower-sign words form a salient class of the American 
braille code. The incidence of these braille shapes, as well as 
that of inkprint sequences which avoid such braille representation, 
deserve careful and analytically explicit study. Nine Grade 2 
braille words define the freimework for the investigation. The text 
sample serving as data base for the observations has been taken in 
principled fashion from the Brown Corpus as described in the second 
paper of this collection and from a recent number of the Readers' 
Digest, and categorized in accordance with Hamp and Caton (1984). 
The occurring categories of representation are defined, and their 
incidence by text subcategory tabulated for frequency and for 
proportionate occurrence. Comparative data is offered on the 
incidence of braille units and inkprint for the letter sequences in 
question. Some initial comment is offered on various aspects of 
the numerical findings, including the implication of importance for 
the learning and teaching of braille. 



85 



The Text Frequency and Incidence 
of Lower-Sign Sequences in American Braille 

Introduction: Scope of the Study 

The lower-sign words of braille form a class which has 
justifiably attracted the attention of scholars and users of 
braille.^ The mere fact that those closest to braille have 
perceived them as a group indicates that they must engage our 
serious attention in a systematic way if we are to study the 
braille code with the care and structured precision which it 
deserves . That is of course a declaration of intellectual 
obligation, a duty which we owe to one aspect of our strivings to 
understand and explain the mechanics of human behavior and 
cognition. There is also a practical side to such a concern which 
is of obvious importance: The attentive study of such a class of 
signs and their surrogates must surely yield in the end 
implications for the sound teaching, effective use, and wise 
administration of the braille code. Valid instruction, viable 
instructional materials,^ and even revisions in the code, envisaged 
or mooted, must be grounded in the observation of relevant data and 
in the analysis and formulation of its systematic properties. The 
analysis must be exact and informed. 

The lower-sign words form a set which may in an obvious way be 
defined as having as members logogreuns characterized by a 
restricted dot configuration in their shapes. The physical and 
perceptual attributes and consequences of this concentration in the 



86 



disposition of their dots have been studied long since by scholars 
of braille.^ It is also clear to any grammarian that the English 
words represented by these lower-sign shapes are for the most part 
highly basic short working elements of English grammar. 

The purpose of the present study is to investigate the 
incidence of occurrence and avoidance of the lower-sign words and 
syllables in normal running American English text. It is intended 
eventually to study the distribution and dependencies of 
occurrence, i.e. the syntax or syntactic properties in the broadest 
sense of these signs and allied sequences in relation to other 
braille elements and to the grammatical syntax of English. As a 
first step in this direction we present here, largely in tabular 
form and with some commentary, the numerical account of the 
incidence of the lower signs and their surrogates in a significant 
sample of various genres of American English prose. 

We must first distinguish the categories of use of those signs 
and their related sequences. 

In Grade 2 braille the words which are designated as lower- 
sign words are nine in number, (be), (by), (enough), (his), (in), 
(into) , (to) , (was) and (were) . These lower-sign words which are 
also logograms, are represented by cells in which only the lower 
two rows are occupied by dots . 

In order to study the occurrence of these words in samples of 
braille text, we selected a corpus consisting of sections taken 
from the Brown Corpus and of articles from the April 1991 issue of 
the i^eader's Digest translated into braille. The sections from the 



87 



Brown Corpus had been selected in the manner and according to the 
principles described by the authors in the second paper of this 
collection. 

The letter sequences which form these English words in print 
may be represented in various ways in braille. In many cases, of 
course, we are concerned with the whole English word. The English 
word may be contracted,^ partly contracted,^ or uncontracted* in its 
braille representation. In some cases, certain of these print 
sequences, which can in themselves be contracted,* may occur as 
part of a longer sequence of inkprint letters for which there is a 
different braille contraction,* itself a logoqreun in the terms of 
Hamp and Caton (1984). Also, these inkprint sequences often 
constitute parts of other words.' Such words may or may not have 
meanings related to the meanings which the sequences have when they 
stand alone. ^ 
Categories of Representation 

From the Sconple text it appeared that there were 18 possible 
different relations between braille representations and the 
inkprint letter sequences of interest to a consideration of lower 
signs in braille. In each case the situation is characterized in 
brief form and in terms that will be fcuniliar to users, teachers, 
and scholars of braille within a long-standing tradition, and the 
categorization of Hamp and Caton (1984) follows in parentheses. 
There is in what follows no essential departure from the current 
practice and system of braille. The novelty lies in the 
completeness of the sample and display and in the consistency of 



88 



the analytic categorization presented. 

1. Whole words contracted such as (be), (his), etc. 
(logograms) 

2. Whole words partly contracted because of being adjacent 
to another lower-sign word. (An example would be the 
partial contracting of (enough) in (to ) (en) (ou) (gh) . 
This category, strictly consisting of phonograms, while 
possible, did not seem to be found in the scunple.) 

3 . Whole words partly contracted because proximity to the 
end of the braille line did not permit annexing the next 
word. for excunple (in) to (phonograms) (alphabetic 
letters) 

4. Whole words partly contracted because of adjacent 
punctuation. ( phonograuns ) (alphabetic letters) 

5. Whole words uncontracted because of being adjacent to 
another lower-sign word. (alphabetic letters) 

6. Whole words uncontracted because proximity to the end of 

the braille line did not permit annexing the next word. 

f 
( alphabetic letters ) 

7. Whole words uncontracted because of adjacent punctuation. 

(alphabetic letters) 

8. Lower signs as parts of contractions and maintaining 
their shape. An excunple is (he) as part of (before) in 
the contraction (be)f, (parts of logograms) 



89 



bn': 



9 . Lower-sign surrogates as parts of contractions 
conventionally viewed as showing the lower sign altered. 
An example is to as part of (today) in the contraction 
td. (parts of logograims) 

10. Inkprint letter sequences as sections of compound words. 
An example is in in drive-in, (alphabetic letters) 

11. Inkprint letter sequences as parts of other words, 
contracted, but unrelated in meaning. Some examples are 
(in) in f(in)d (phonograms) or (in) in (in) adequate 
(morphograms) . 

12. Inkprint letter sequences as parts of other words, 
contracted, and related in meaning. An example is (be) 
in (he) (ing) (phonograms) or (in) in (in) side 
(morphograms) . 

13. Inkprint letter sequences as parts of other words, 
uncontracted, and unrelated in meaning. An example is by 
in baby. (alphabetic letters) 

14. Inkprint letter sequences as parts of other words, 
uncontracted, and related in meaning. An exeimple is to 
in onto, (alphabetic letters) 

15. Inkprint letter sequences as parts of other words, in 
which one or more letters of the lower-sign letter 
sequence are combined with adjacent letters in 
contractions, and the meaning of the word is unrelated to 
the meaning of the corresponding lower- sign word. 



90 



Examples are his in (wh)isp(er) , was in wa(sh) (ed) , and 
to in a(st) (ound) (ing) . (phonograms) (morphograms) 

(alphabetic letters) 

16. Inkprint letter sequences as parts of other words, in 
which one or more letters of the lower-sign letter 
sequence are combined with adjacent letters in 
contractions, and the meaning of the word is related to 
the meaning of the corresponding lower-sign word. An 
example is to in t(ow)(ar)d. ( phonogrcims ) 

17. Inkprint letter sequences as parts of other words, partly 
contracted, and unrelated in meaning. An example is 
(in) to in m(ount)a(in)top. (phonograms) (alphabetic 
letters) (morphograms) 

18. Inkprint letter sequences as parts of other words partly 
contracted, and related in meaning. Some examples are 
w(er) (en) ^t and (there) (in)to, (phonograms) (alphabetic 
letters) (morphograms) 

In some cases a lower sign occurrence may belong to two 
categories. Examples: in (to )be, the sequence be is adjacent to 
a lower sign and also to punctuation and therefore falls into 
categories 5 and 7; in 1 (ow) (er) (ed) the were sequence is partly 
contracted and letters from the sequence combine with adjacent 
letters to form contractions so that this occurrence belongs to 
categories 15 and 17. 



91 



Summary of Categories of Braille Usage of Lower-Sign Sequences 

1. Whole word, contracted 

2. Whole word, partly contracted, adjacent to lower-sign 
word 

-i 3. Whole word, partly contracted, at end of braille line 

4. Whole word, partly contracted, adjacent to punctuation 

5. Whole word, uncontracted, adjacent to lower sign word 

6. Whole word, uncontracted, at end of braille line 

7. Whole word, uncontracted, adjacent to punctuation 

8. Part of another contraction, shape retained 

9. Part of another contraction, shape altered 

10. Section of a compound word 

11. Part word, contracted, unrelated meaning 

12. Part word, contracted, related meaning 

13. Part word, uncontracted, unrelated meaning 

14. Part word, uncontracted, related meaning 

15. Part word, overlapped contractions, unrelated meaning 

16. Part word, overlapped contractions, related meaning 

17. Part word, partly contracted, unrelated meaning 

18. Part word, partly contracted, related meaning 

We try in this study to dwell on central issues and 
characteristics of the topic. Because we attempted to cast as wide 
a net as possible three types of phenomena have appeared in our 
data which seem to us peripheral in varying degrees; they result 
from the attempt to make complete correlations with ink- print 
spelling. 



92 



1. Probably categories 13 and 14 are of less central 
interest to our topic than many others, but they are 
interesting in exploring the terrain (e.g. hiss(ing) ) and 
defining the limits, and they are probably of more 
concern than the tricky issues of 15, 16, and 17, which 
involve debatable theoretical questions of analysis. 

2. Categories 15, 16, and 17 forced themselves on our 
attention because of our insistence on inspecting all 
sequences which qualified under any definition. Many of 
the examples here could be eliminated on other grounds 
and we may wish later to prune our list; but for the 
present we feel it is better to allow our lists to stand. 

3 . There are a few items which have surfaced under 
categories 9, 15, and 16 which our procedures threw up 
(e.g. (this) f wa(sh) (ed) , (wh)i(st)le, t(ow) (ar)d) and 
which to varying degrees we would find inconsistent with 
our analytic principles for the braille code. For the 
present we allow them to stand, since they do not distort 
our main results. 

Brown Corpus Occurrences 

The tabulations which follow show the niimerical occurrence in 
the sample text corpus (Brown Corpus) of the lower-sign sequences 
which have been identified above subcategorized by the above 18 
syntagmatic categories (in columns) and by genre-specified (A-R) 
text corpus source (in rows) . Each tabulation is identified and 
titled by its relevant lower-sign word; but it should be noted 



93 



clearly that all 18 categories of sequence are accounted for. 

The Brovm Corpus literature types, or genres, associated with 
the code letters in the first, identifying, column are: A. Press: 
Reportage; B. Press: Editorial; C. Press: Reviews; D. 
Religion; E. Skills and Hobbies; F. Popular Lore; G. Belles 
Lettres, Biography, etc.; H. Miscellaneous; J. Learned and 
Scientific Writings; K. Fiction: General; L. Fiction: Mystery 
and Detective; N. Fiction: Adventure and Western; P. Fiction: 
Romance and Love Story; and R. Humor. 



94 



Brown Corpus - Samples 



10 11 12 13 14 15 16 17 18 



AOl 


20 


A21 


6 


BOl 


15 


COl 


2 


DOl 


7 


EOl 


5 


E21 


33 


FOl 


5 


F21 


9 


GOl 


8 


G21 


6 


G41 


3 


G61 


5 


HOI 


11 


H21 


13 


JOl 


16 


J21 


15 


J41 


13 


J61 


9 


KOI 


2 


K21 


5 


LOl 


2 


NOl 


4 


POl 


6 


ROl 


5 



1 


- 


- 


2 


- 


— 


4 


5 


2 


7 


14 


- 


- 


- 


6 


- 


- 


4 


3 


5 


14 


11 


2 


- 


- 


7 


- 


- 


9 


1 


4 


9 


9 


2 


- 


- 


2 


- 


- 


3 


3 


9 


6 


8 


2 


- 


- 


6 


- 


- 


5 


4 


6 


8 


3 


1 


- 


- 


18 


- 


- 


4 


- 


14 


2 


13 


4 


- 


1 


3 


- 


- 


3 


1 


4 


2 


6 


- 


- 


1 


4 


- 


- 


6 


- 


3 


4 


8 


2 


- 


- 


5 


- 


- 


3 


- 


1 


8 


12 


5 


- 


- 


9 


- 


- 


7 


3 


6 


16 


22 


3 


- 


1 


8 


- 


- 


4 


- 


4 


9 


25 


6 


- 


- 


6 


- 


- 


2 


- 


- 


4 


8 


4 


- 


1 


8 


- 


- 


8 


- 


2 


9 


10 


3 


- 


- 


1 


- 


- 


- 


- 


1 


1 


4 


7 


- 


- 


4 


- 


- 


3 


2 


2 


10 


15 


5 


- 


- 


6 


- 


- 


3 


- 


4 


12 


12 


1 


- 


1 


1 


- 


- 


- 


1 


2 


1 


2 


1 


- 


- 


8 


- 


- 


7 


- 


3 


5 


4 


8 


- 


- 


4 


- 


- 


7 


2 


- 


6 


9 


2 


- 


- 


3 


- 


- 


1 


- 


6 


2 


23 


3 


- 


1 


7 


- 


- 


6 


1 


13 


7 


8 


5 


- 


1 


6 


- 


- 


8 


- 


4 


6 


6 


1 


- 


- 


6 


- 


- 


- 


1 


6 


4 


6 


3 


- 


- 


18 
95 


- 


- 


5 


- 


7 


5 


6 


5 


— 


— 


6 


— 


- 


6 


2 


9 


8 


3 



BY 



Brown Corpus - Samples 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 



AOl 9--__i______2-- 

A21 i5----i---_--__- 

BOl ii___-----__-_i_ 

COl 15____2------1-- 

DOl 5--_-i2-------- 

EOl 7-__-i_______-_ 

E21 7___-_-__-_-_3_ 

FOl 16 -------------- 

F21 20 --------1----- 

GOl 9------_-____-- 

G21 ii__-_i___--_i_- 

G41 7-____-_-__--_- 

G61 7--_-i6-------- 

HOl ii-___22------l- 

H21 19 ----1--------- 

JOl i4-_--i---_----- 

J21 7----21-------- 

J41 16 ----11-------- 

J61 ii-_--i_--i___-- 

KOl 5-_-____--__--- 

K21 5________-___-- 

Loi ___-_--_____i-2 

NOl 2-----1-------- 

POl 6-------------- 

96 

ROl 7____i_--i--2-- 



Brown Corpus - Samples 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 

AOl __-_--_---------- 

A21 __-____---_------. 

BOl ___i_-_---_----_- 

COl _________--____-- 

DOl i_____------_---- 

EOl _______-_____---- 

E21 i__-_-_---------- 

FOl ___i____-____--_- 

F21 2--1------------- 

GOl ____-________-_-- 

G21 _-_i-_-___---_--- 

G41 i--_---___--___-_ 

G61 __--____-_____--- 

HOI I------ __-_-___-- 

H21 -_-_--«_-_------- 

jOl i-__ _____________ 

J21 --____________--- 

J41 _________________ 

J61 5______________-- 

KOl i__i_____________ 

K21 i--i__________-_- 

LOl 2---------------- 

NOl i__________-____- 

POI 2--1------------- 

97 

ROl _-_i____---___--- 



HIS 



AOl 


12 


A21 


11 


BOl 


11 


COl 


12 


DOl 


7 


EOl 


16 


E21 


- 


FOl 


12 


F21 


- 


GOl 


5 


G21 


16 


G41 


9 


061 


4 


HOI 


2 


H21 


- 


JOl 


- 


J21 


- 


J41 


- 


J61 


11 


KOI 


58 


K21 


41 


LOl 


3 


NOl 


21 


POl 


29 


ROl 


30 



Brown Corpus - Samples 
i 6 7 8 9 10 11 12 13 14 15 16 17 

3---3------- 

----10-----3- 
l-_-13----__. _ 

----20-----1- 
1---19-----2- 
----6-----3- 
1---22-----3- 

----18-----3- 
1---18-----2- 

1---12-----9- 

----19-----3- 

__- 28 ------- 

--1-9-----3- 

3---3-----2- 
- 12-----1- 

1--- 11 ------- 

2---3-----1- 

98 
3---2-----2- 



Brown Corpus - Samples 



IN 



AOl 


39 


A21 


51 


BOl 


44 


COl 


33 


DOl 


60 


EOl 


30 


E21 


27 


FOl 


39 


F21 


58 


GOl 


46 


G21 


34 


G41 


33 


G61 


60 


HOI 


36 


H21 


69 


JOl 


37 


J21 


44 


J41 


50 


J61 


56 


KOI 


35 


K21 


25 


LOl 


24 


NOl 


33 


POl 


33 


ROl 


33 



8 9 10 11 12 13 14 15 16 17 18 

13 41-47 2------ 

-2 63 148 5--1--- 

12 66-67 6----- 

- 3 72 - 69 5 ----- - 

-9 44-43 5----- 

-5 58-59 6----- 

2- 60 -67 6----- 

1 100 1 71 10 - - - - - 

- 4 48 - 45 10 - - - - - 
-149-626--3-- 
22 23 -51 6----- 
2 - 50 - 42 2 - - 1 

1 2 39 - 84 15 - - 1 - - 

- 56 - 107 17 - - 58 

-2 65-66 7--!-- 

1- 29 -67 2----- 

- 13 36 - 172 18 

- 55 - 137 4 - - - - - 

- 3 84 - 66 2 - - - - 

3 54 -61 - - - - -- 

-9 83-33 1----- 

-3 62-40 1--1-- 

39 64-39------ 

15 90 -43 4----- 

99 
-5 56 157 1----- 



' INTO 



Brown Corpus - Samples 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



AOl 3-----__-__---___ 

A21 2---------------- 

BOl 2---------------- 

COl 3---_---_________ 

DOl 9_------_____-_-_ 

EOl 5_-______________ 

E21 -__-__-_____--_-_ 

FOl i--i-____-______- 

F21 4----__--___--_-- 

GOl i________________ 

G21 2---------------- 

G41 _________________ 

G61 2-1-------------- 

HOl _________________ 

H21 2---------------- 

jOl __-___-________-i 

J21 13 _-----_______--- 

J41 _________________ 

J61 3-i_-----_--_---2 

KOI 3_-__--_______--- 

K21 9----____->_----- 

LOl 3---____-____---- 

NOl 9__-____--_------ 

POI 5_______-_____--- 

100 

ROl 5-____---___----- 



Brown Corpus - Samples 



TO 



10 11 12 13 14 15 16 17 18 



AOl 48 

A21 36 

BOl 42 

COl 41 

DOl 42 

EOl 39 

E21 42 

FOl 47 

F21 31 

GOl 45 

G21 55 

G41 70 

061 35 

HOI 49 

H21 32 

JOl 28 

J21 13 

J41 50 

J61 75 

KOI 41 

K21 57 

LOl 71 

NOl 45 

POl 42 

ROl 65 



6 


2 


3 


1 


- 


- 


- 


40 


- 


3 


1 


4 


- 


2 


15 


1 


- 


- 


14 


- 


6 


1 


6 


- 


2 


3 


1 


- 


- 


19 


- 


8 


5 


1 


1 


3 


- 


- 


- 


- 


18 


- 


5 


- 


4 


- 


9 


1 


- 


- 


- 


5 


- 


2 


- 


3 


2 


5 


- 


1 


- 


- 


19 


- 


5 


- 


6 


- 


- 


1 


- 


- 


- 


23 


- 


7 


- 


2 


- 


1 


1 


- 


- 


- 


6 


1 


9 


- 


3 


- 


4 


1 


- 


- 


- 


8 


- 


3 


- 


6 


1 


1 


3 


1 


- 


- 


8 


- 


6 


2 


7 


- 


2 


- 


- 


- 


- 


15 


- 


3 


1 


7 


- 


- 


- 


- 


- 


- 


12 


- 


7 


- 


1 


- 


2 


1 


- 


- 


- 


7 


1 


10 


2 


7 


- 


- 


1 


- 


- 


- 


20 


- 


2 


1 


4 


- 


2 


2 


2 


- 


- 


11 


- 


5 


- 


4 


3 


- 


1 


- 


- 


- 


16 


- 


3 


- 


3 


10 


13 


- 


- 


- 


- 


19 


- 


- 


- 


8 


1 


- 


- 


- 


- 


- 


4 


- 


3 


- 


6 


1 


3 


3 


- 


- 


- 


19 


1 


13 


- 


4 


3 


3 


1 


- 


- 


- 


18 


- 


5 


1 


5 


- 


9 


- 


- 


- 


- 


10 


- 


14 


3 


2 


1 


3 


1 


- 


- 


- 


12 


- 


9 


1 


- 


- 


9 


2 


- 


- 


- 


11 


- 


8 


4 


3 


- 


5 
101 


- 


- 


- 


- 


13 


1 


20 


2 


4 


— 


5 


1 


— 


— 


- 


11 


— 


9 


2 



WAS 



Brown Corpus - Samples 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



AOl 18 ------------11 

A21 28 ------------1- 

BOl i4__--_-----__-2 

COl 8-------------1 

DOl 5--___--_______ 

EOl 9______--_-_-i_ 

E21 _-___--_-_____i 

FOl 12 -------------1 

F21 ii_-___--__---_i 

GOl 4--_-___-___--- 

G21 22 -----1-------- 

G41 53--_-___---__-3 

G61 23 -------------- 

HOI __--___-_>-_-- 11 

H21 7--___---_----- 

JOl i4-----____-_-_i 

J21 _______________ 

J41 _______________ 

J61 30-----1-------6 

KOI 37_____i______-- 

K21 21------------ 11 

LOl 28 ------------1- 

NOl 32 ------------3 1 

POl 36 -------------- 

102 

ROl 39_____2------l- 



WERE 



Brown Corpus - Samples 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



AOl 2---------------- 

A21 17 --_-_-_---_--_-- 

BOl 2-------------1-1 

COl _-______-___-__-_ 

DOl 3__--_--_______-_ 

EOl i_--_---__-_-___- 

E21 _____-----___--__ 

F013--1----------1-1 

F21 6---------------- 

GOl i--_--_---_-_-_-- 

G21 2-------- -------- 

G41 12------------ - _ _ - 

G61 10 ---------------- 

HOI i____----_____-_- 

H21 __--___-__>__-___ 

JOI 2---------------- 

J21 i--__--________-- 

J41 ---_____________- 

J61 i3_-i-----__------ 

KOl 7----__-_____-i-l 

K21 5-____-_____--2-2 

LOl ii-_i-_-_______i-l 

NOl i3_____________2-2 

POl 5_---______-__--- 

103 

ROl 7-__-________---- 



Occurrences in Reader's Digest 

Shown next are tabulations of the occurrences of the lower- 
sign sequences in our Reader's Digest text. The tabulations again 
show the numerical occurrence in the Sconple text corpus (i?eader's 
Digest) of the lower-sign sequences which have been identified 
above subctegorized by the above 18 syntagmatic categories (in 
columns) . Each tabulation is identified and titled by its relevant 
lower-sign word; but it should be noted clearly that all 18 
categories of sequence are accounted for. Magazine articles are 
numbered from 1 through 36, the first page of each article being 
given in the next left-hand column. The 37th section consists of 
humorous items found throughout the magazine, the first of which is 
on page 16 . 



104 



Reader's Digest April, 1991 



BE 



10 11 12 13 14 15 16 17 18 



1 


2 


7 


2 


13 


2 


3 


21 


2 


4 


27 


5 


5 


37 


3 


6 


47 


1 


7 


49 


4 


8 


51 


5 


9 


59 


1 


10 


63 


1 


11 


69 


5 


12 


73 


5 


13 


79 


4 


14 


84 


2 


15 


91 


1 


16 


93 


5 


17 


98 


2 


18 


105 


6 


19 


109 


6 


20 


116 


1 


21 


122 


2 


22 


125 


- 


23 


129 


4 


24 


134 


9 


25 


137 


8 


26 


142 


2 


27 


145 


1 


28 


151 


1 


29 


155 


- 


30 


159 


6 


31 


167 


1 


32 


169 


- 


33 


171 


4 


34 


179 


3 


35 


183 


24 


36 


200 


_ 


37 


16 


19 



1-12 
- - - 4 
2 - - 4 
6 - - 1 



2 - - 8 
1 - - 1 
5 - - 2 



3 - - 2 

- - 15 

3 - - 3 

1 - - 5 



3 - - 3 
2 - - 6 
1-16 

4 - - 6 

1 - - 8 

2 - - 2 
1 - - 1 

- - 2 

1 - - 1 
- - - 6 

2 - - 3 

2 - - 4 

2 - - 6 

2 

3 - - 9 



- - - 2 

- - - 4 
16 - 1 35 

- - - 1 
5 - - 11 



105 



2 


2 


1 


- 


3 


1 


- 


3 


2 


19 


- 


- 


2 


1 


7 


3 


1 


3 


1 


6 


5 


- 


1 


1 


7 


2 


1 


2 


4 


9 


1 


- 


- 


- 


2 


14 


1 


4 


7 


7 


1 


- 


11 


3 


8 


3 


1 


13 


7 


9 


1 


1 


2 


1 


3 


3 


2 


4 


19 


19 


4 


- 


4 


1 


5 


10 


2 


4 


8 


18 


1 


- 


1 


- 


1 


8 


1 


3 


6 


11 


7 


1 


2 


2 


12 


3 


3 


1 


- 


2 


2 


1 


4 


6 


10 


2 


1 


2 


2 


2 


3 
1 
4 


1 


1 


3 


1 


_ 


4 


7 


9 


8 


- 


2 


2 


5 


2 


1 


1 


9 


5 


5 


1 


1 


3 


2 


- 


2 


9 


- 


6 


1 


- 


- 


2 


- 


- 


- 


6 


- 


2 


4 


1 


3 


2 


13 


1 


_ 


4 


4 


3 


4 


- 


1 


1 


- 


3 


- 


4 


2 


1 


1 


1 


1 


1 


6 


36 


19 


15 


32 


42 


_ 


_ 


1 


— 


2 


14 


6 


20 


13 


30 



Reader's Digest April, 1991 



BY 



1 


2 


4 


2 


13 


2 


3 


21 


7 


4 


27 


3 


5 


37 


3 


6 


47 


2 


7 


49 


4 


8 


51 


13 


9 


59 


8 


10 


63 


3 


11 


69 


3 


12 


73 


6 


13 


79 


5 


14 


84 


18 


15 


91 


3 


16 


93 


8 


17 


98 


11 


18 


105 


1 


19 


109 


7 


20 


116 


8 


21 


122 


6 


22 


125 


1 


23 


129 


6 


24 


134 


4 


25 


137 


7 


26 


142 


6 


27 


145 


9 


28 


151 


4 


29 


155 


2 


30 


159 


4 


31 


167 


2 


32 


169 


2 


33 


171 


5 


34 


179 


2 


35 


183 


2 7 


36 


200 


2 


37 


16 


62 



10 11 12 13 14 15 16 17 ] 



14 



2 - - - 1 
2 - - - 1 



1------1- 

21------1 

-1-----4- 

1---1--21 



1 

2 



1 



24--1--10 2 



106 



Reader's Digest April, 1991 



ENOUGH 



1 2 - - - - 

2 13 1 - - - 

3 21 - - - - 

4 27 - - - - 

5 37 - - - - 

6 47 - - - - 

7 49 - - - - 

8 51 1 - - - 

9 59 - - - - 

10 63 - - - - 

11 69 - - - - 

12 73 1 - - - 

13 79 - - - - 

14 84 - - - - 

15 91 - - - - 

16 93 1 - - - 

17 98 - - - - 

18 105 

19 109 2 - - - 

20 116 _ - - _ 

21 122 

22 125 

23 129 2 - - - 

24 134 

25 137 - _ - - 

26 142 2 

27 145 

28 151 _ - _ - 
2 9155 

30 159 - - - - 

31 167 1 - - - 

32 169 - - - - 

33 171 

34 179 _ - - - 

35 183 4 - _ _ 

36 200 _ _ _ _ 

37 16 - - - 3 



10 11 12 13 14 15 16 17 18 



107 



Reader's Digest April, 1991 



HIS 



1 


2 


2 


2 


13 


2 


3 


21 


5 


4 


27 


5 


5 


37 


2 


6 


47 


4 


7 


49 


- 


8 


51 


3 


9 


59 


7 


10 


63 


5 


11 


69 


18 


12 


73 


20 


13 


79 


8 


14 


84 


22 


15 


91 


3 


16 


93 


3 


17 


98 


6 


18 


105 


5 


19 


109 


28 


20 


116 


- 


21 


122 


2 


22 


125 


6 


23 


129 


13 


24 


134 


2 


25 


137 


6 


26 


142 


14 


27 


145 


- 


28 


151 


5 


29 


155 


- 


30 


159 


5 


31 


167 


1 


32 


169 


2 


33 


171 


- 


34 


179 


12 


35 


183 


22 


36 


200 


1 


37 


16 


48 



10 11 12 13 14 15 16 17 



2 ----- 1 
8 ----- 3 



2---7-----1 

3 ----- 2 
2---2------ 

1---3-----1 
1---3-----3 

1---1---1-1 

1---7-----.1 
1---9------ 

1-1-1------ 

1_--4_____1 



5 ----- - 

5 ----- 2 

3 ----- 1 

28 ----- 2 



- 25 



108 



Reader's Digest April, 1991 



IN 



1 


2 


26 


2 


13 


20 


3 


21 


16 


4 


27 


18 


5 


37 


32 


6 


47 


11 


7 


49 


15 


8 


51 


49 


9 


59 


28 


10 


63 


27 


11 


69 


36 


12 


73 


45 


13 


79 


34 


14 


84 


52 


15 


91 


6 


16 


93 


27 


17 


98 


32 


18 


105 


21 


19 


109 


52 


20 


116 


22 


21 


122 


22 


22 


125 


8 


23 


129 


56 


24 


134 


12 


25 


137 


35 


26 


142 


26 


27 


145 


22 


28 


151 


9 


29 


155 


7 


30 


159 


10 


31 


167 


8 


32 


169 


3 


33 


171 


20 


34 


179 


19 


35 


183 


112 


36 


200 


2 


37 


16 


112 



10 11 12 13 14 15 16 17 18 



- 


1 


25 


- 


26 


2 


- 


2 


28 


- 


22 


1 


- 


1 


34 


- 


30 


2 


1 


2 


43 


- 


28 


- 


- 


7 


59 


1 


77 


4 


_ 


2 


31 


_ 


20 


_ 


- 


2 


20 


- 


12 


3 


1 


12 


129 


2 


156 


17 


- 


1 


49 


- 


32 


3 


- 


4 


51 


- 


26 


6 


— 


— 


40 


1 


29 


— 


- 


3 


78 


- 


39 


3 


2 


3 


62 


- 


37 


5 


1 


15 


97 


- 


60 


7 


1 


2 


30 


- 


26 


4 


— 


2 


61 


— 


110 


6 


- 


4 


66 


- 


48 


2 


2 


2 


47 


- 


55 


3 


- 


4 


56 


- 


49 


9 


1 


2 


105 


- 


49 


5 


— 


1 


27 


_ 


32 


1 


1 


2 


11 


- 


10 


1 


- 


3 


58 


- 


48 


3 


- 


1 


39 


- 


26 


3 


— 


3 


89 


- 


91 


4 


— 


- 


38 


— 


19 


3 


- 


2 


31 


- 


23 


1 


- 


3 


19 


- 


15 


1 


- 


- 


8 


- 


11 


- 


1 


- 


34 


3 


29 


1 


— 


4 


29 


_ 


20 


2 


- 


1 


19 


- 


9 


2 


1 


1 


28 


- 


32 


6 


- 


3 


38 


- 


23 


2 


2 


37 


416 


- 


218 


9 


— 


1 


10 


— 


11 


4 


4 


11 


244 


2 


203 


11 



109 



Reader's Digest April, 1991 



INTO 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



121---------------- 

2 13 2---------------- 

3 21 i---------------- 

4 27 2---------------- 

5 37 7---------------- 

6 47 2---------------- 

7 49 2---------------- 

8 51 12 ---------------- 

9 59 1---------------- 

10 63 4-1-------------- 

11 69 ----------------- 

12 73 3---------------- 

13 79 3---------------- 

14 84 15 ---------------- 

15 912---------------- 

16 93 2---------------- 

17 98 4---------------- 

18 105 2------------- 1-1 

19 109 4---------------- 

20 116 2-2----- --------- 

21 122 1---------------- 

22 125 2---------------- 

23 129 3---------------- 

24 134 1---------------- 

25 137 3---------------2 

26 142 ----------------- 

27 145 2---------------- 

28 151 3-1-------------- 

29 155 ----------------- 

30 159 ----------------- 

31 167 4-1-------------- 

32 169 i-i_---_-_------- 

33 171 1---------------- 

34 179 3---------------- 

35 183 37-1-------------1 

36 200 1---------------- 

37 16 11 ---------------1 

110 



Reader's Digest April, 1991 



TO 



1 


2 


11 


2 


13 


29 


3 


21 


15 


4 


27 


31 


5 


37 


39 


6 


47 


16 


7 


49 


12 


8 


51 


60 


9 


59 


28 


10 


63 


37 


11 


69 


45 


12 


73 


39 


13 


79 


50 


14 


84 


60 


15 


91 


7 


16 


93 


31 


17 


98 


46 


18 


105 


34 


19 


109 


39 


20 


116 


48 


21 


122 


21 


22 


125 


10 


23 


129 


50 


24 


134 


46 


25 


137 


37 


26 


142 


28 


27 


145 


22 


28 


151 


15 


29 


155 


4 


30 


159 


30 


31 


167 


12 


32 


169 


14 


33 


171 


18 


34 


179 


37 


35 


183 


275 


36 


200 


8 


37 


16 


172 



10 11 12 13 14 15 16 17 18 



2 


- 


1 


1 


- 


1 


2 


- 


3 


- 


1 


- 


3 


1 


2 


1 


5 


- 


7 


- 


2 


- 


2 
2 

12 


- 


5 


1 


1 


3 


- 


1 


2 


1 


- 


4 


1 


2 


_ 


— 


1 


2 


- 


3 


2 


3 


1 


3 


2 


4 


1 


15 


- 


1 


- 


2 


- 


3 


— 


2 


— 


4 


- 


4 


2 


2 


1 


2 


- 


7 


- 


4 


3 


- 


- 


2 


- 


2 


- 


1 
2 
3 


1 


6 


— 


_ 


1 


- 


1 


1 


1 


1 


3 


1 


2 


— 


— 


3 


1 
1 


- 


2 

3 


- 


3 
2 


— 


4 
1 

1 


— 


2 


— 


— 


4 


- 


3 


2 


25 


1 


37 


4 


— 


— 


1 


2 


13 


- 


11 


5 



11 


1 


6 


1 


8 


- 


10 


- 


14 


- 


3 


1 


10 


- 


2 


- 


28 


- 


4 


1 


7 


- 


2 
2 

7 


- 


34 


_ 


2 


11 


- 


4 


1 


10 


1 


7 


- 


10 


— 


13 


— 


24 


- 


6 


- 


17 


- 


9 


1 


33 


1 


13 


1 


2 


- 


1 


- 


62 


_ 


5 


_ 


18 


- 


10 


1 


26 


- 


2 


- 


20 


- 


11 


- 


11 


4 


13 


3 


3 


1 


1 


- 


10 


- 


5 


- 


11 


- 


8 


1 


11 


- 


1 


1 


12 


- 


5 


— 


12 


— 


10 


- 


15 


- 


2 


1 


2 


2 


1 


1 


- 


- 


1 


2 


7 


- 


3 


1 


2 


1 


14 


— 


4 


1 


- 


- 


13 


- 


2 


- 


8 


- 


- 


- 


80 


5 


41 


1 


1 


— 


— 


1 


66 


5 


33 


- 



Ill 



Reader's Digest April, 1991 



WAS 



10 11 12 13 14 15 16 17 18 



123-------------- 

2 13 2------------1- 

3 217-------------3 

4 27 13 -------------- 

5 37 17 ------------13 

6 47 7-------------2 

7 49 5-------------- 

8 5145-----2------43 

9 59 17-----1------11 

10 63 31 ------------3 3 

11 69 32 ---_--_-_--- 1 

12 73 41 -------------2 

13 79 34-----1------1- 

14 84 26 -------------- 

15 91 1-------------- 

16 93 25 -------------- 

17 98 19 -------------2 

18 105 3-----1-------2 

19 109 26 -------------- 

20 116 32 -------------- 

21 122 4-------------- 

22 125 14 -----2------ 1- 

23 129 13 -------------2 

24 134 5------------1- 

25 137 2-------------4 

26 142 19 -------------- 

27 145 --------------- 

28 151 6-------------- 

29 155 5 -- -- __-_------ 

30 159 1-------------- 

31 167 15 ------------2- 

32 169 4------------- 1 

33 171 2 -- -- ---------- 

34 179 14-----1-------- 

35 183 186 -----3------22 

36 200 2-------------- 

37 16 70 -----3------ 13 

112 



Reader's Digest April, 1991 



WERE 



10 11 12 13 14 15 16 17 18 



122----------------- 

2 13 1----------------- 

3 21 ------------------ 

4 27 5-------------1-1- 

5 37 3-------------1-1- 

6 47 2----------------- 

7 49 3----------------- 

8 51 10 -------------2-2- 

9 59 5----------------- 

10 63 8-------------1-1- 

1169 4-------------1-1- 

12 73 7-------------1-1- 

13 79 5--------------1-1 

14 84 27 ----------------- 

15 91 ------------------ 

16 93 9--------------1-1 

17 98 11 ----------------- 

18 105 ------------------ 

19 109 7----------------- 

20 116 5----------------- 

21 122 2----------------- 

22 125 1-------------1-1- 

23 129 3------------- 2 -2- 

24 134 1----------------- 

25 137 2----------------- 

26 142 2----------------- 

27 145 --------------1-1- 

28 151 1----------------- 

29 155 ------------------ 

30 159 2----------------- 

31167 1----------------- 

32 169 --------------1-1- 

33 171 ----__-_____--_--- 

34 179 2----------------- 

35 183 42--1-----------1- 

36 200 ----_--_-____-_--- 

37 16 31--1----------616: 

113 



Devia-bions 

By a quick scanning with the eye a few deviations will 
immediately be noticed in the Brown Corpus (by genre) and in the 
Reader's Digest (by magazine article), for example, for (in) 
(category 15) and for (by) (category 13). These are doubtless to 
be credited to idiosyncracies in lexical choice or topic of the 
text or genre. A closer and refined inspection of such disparities 
would no doubt give interesting results bearing on braille usage in 
relation to English discourse and style. 
Proportionate Occurrence 

In the next tables occurrences of the lower-sign sequences are 
summarized according to the types of text in which they appear. 
The letters in the first column are those which identify the 
classifications of material in the Brown Corpus based on subject 
and style. The percentage distribution across the 18 categories of 
braille representation is given for each of the lower-sign 
sequences in each of the text types and in the Brown Corpus samples 
as a whole. By way of comparison, for each 

lower-sign sequence, total and percentage figures are given for the 
sequence's occurrences in the i?eader's Digest text. 



114 



^r 


»- 




<o 




N 


t- 


t* 


m* 




eo 


»» 


^ 


CM 


P» 


o 


CM 


■* 


T 


u> o 


a> o 


eo 


eo 


«o 


« 


0> <0 


O -H 


in eo 


a> «o 


e- o 


»4 V 


(D CM 


(O V 


CO o 


« eo 


t» o> 


(«■ CM 


w • 


• 




■ 




• 


w4 • 


C4 • 


(O . 


•-< 


• 


CM • 


« • 


• 


• 


• 


• 


^ • 


eo • 


^ 


(O 




N 




t- 


(O 


eo 


o 




'B' 


«8 


^ 


«D 


^4 


CM 


(D 


CM »< 


CM V 


M 


*^ 




N 






•<i4 


CM 


c» 




CM 


>4 


CO 


-* 


CM 


^4 




CM 


CM 


o> 


t» 




'»• 




«»< 


v4 


o 


w« 




o> 


o> 


o 


CM 


0> 


O 


eo 


in 


^ 


»4 v4 


o» o 


(O 


•4 


eo lo 


^ US 


««» o» 


eo o 


•4 


CM 


T CM 


o> o 


(D CM 


•<» CM 


leo 


eo -4 


in <o 


CM eo 


M • 






• 




• 


• 


•H • 


n - 


•■4 


• 


CM • 


• 


• 


• 


• 


• 


(O . 


tn • 


O 


«o 




t<- 




O) 


« 


to 


eo 

Vi4 




w4 




o 


«o 

w4 


^ 
«« 


o 

•i4 


00 


•i4 


»4 


r» 


^ 




^ 




« 


e> 


n 


o> 




O 


<D 


«4 


w4 


« 


O 


lO 


a> 


CM 


e- t- 


-V -4 


o> t*' 


to to 


eo e- 


^ o 


CM <D 


eo 


o> 


a> « 


o» -< 


^ eo 


«D ^ 


*>- o 


o» •«»• 


»«• « 


o eo 


• 


• 




• 




• 


Wi4 • 


• 


^4 • 




• 


• 


•M • 


• 


• 


• 


• 


•H • 


^ • 


<o 


e- 




lO 




'T 


IS 


lO 


lO 




« 


to 


«4 


o 


V4 


^ 


o 


— o 


•>« —4 








C4 




«M 


^4 












CM 


^■4 


CM 


•^ 


CM 


«M 


-« 


e» 


o> 




t- 




(D 


eo 




CM 




o 


a> 


«M 




t- 




to 


eo 


CM 


00 «D 


-* t- 


« lO 


»* 


c~ 


-< oo 




C^ ^ 


CM «D 


« t- 


«4 ^ 




,«4ia 




CM *r> 


o> in 


O CM 


• 


• 




• 




• • 


• 




• 




* 


• 


• 




• 




m 


CM • 


in • 


t- 


v4 




00 




a> 


' 




^4 




CM 


V4 


•H 




« 




^ 


CM 


^ 


a> 


t- 




t- 




o 


^ 


eo 


to 




o 


M 


CO 


CM 




o 


^ 


at 


•4 


eo <o 


a> o 


« 


*a 


lA 


CM 


r*- ^ 


o> «o 


-^a> 


« o> 


t- ^ 


e- t- 


eo (c 




lO o 


«o to 


00 to 


o in 


• 


• 




• 




• 


• 


• 


CM • 




• 


«M • 


• 


* 




* 


• 


o . 


«D • 


t- 


to 




eo 




M 


(O 


CM 


o> 




p» 


O 


t- 


«4 




o 


n 


— o> 


•4 eo 



00 (O 



o 
in 



eo 


CM 




00 


o> 


O) 


CO to 


»< ^r 


a> 


«D 


«4 (S • 


to V 


• 


CM - 




• 


eo • 


• 


■V 


eo 




CM 


^ 


to 


VN 


«-4 




«i4 


«-< 






eo 




•H 


« to 






•H eo 


»4 


V 


CM a> 





o — 



«4 (D •H ' 



CD CM 



« O 



o 


^ 


00 


00 o 


<D (D 


"* w 


fH • 


• 


to • 


«o 


eo 


-« eo 


eo 


»H 


« CM 



e> CM 
«D • 



« to 
eo CM 



■O •H 



CM m 



eo 

CM 00 



a> 

to CO 



CM 
CM 00 



« eo 

00 to 






CO 
CD 



to to 



to le 
eo 



o 
eo o 



in CO 



•» in 
«o v> 



« e 
-« o 



o 


at 


«o o 


in t- 


CM • 


^4 « 


in 


<D 


e« 


CM 



e~ o 



CO 


CM 


CO 


t~ 


to 


00 


eo eo 


V r- 


CM •» 


^ -^ 


CO tn 


e» t- 


CO • 


•-< • 


CM • 


CM • 


tn • 


• 


CO 


A 


o 


»H 


«4 


t* 


CO 


«>« 


«^ 


o 


CO 





o> 


o 


(O 


00 


c« 


(D O 


in eo 


tn o> 


• 


• 


• 


CM • 


^ 


CM 


«4 


CM 0> 



<o 
in 



CD 



00 

to 



115 



o 



CO 
CM 



o 
to 



(O 
CM 



h 
O 

. . to 

t» e- • 

*» 

X X « 

b U U 
O O 

w to e 
a> «> >H 

ts « o 
u o ■ 

e e « 

o 

O O b 

a ■ « 
■ em 



B ■ X 

o 
into «o 

« 
X X.*' 
b b « 
O O U 

to to 
« V e 

4J 4J <^ 

cd « 

CUB 
V 

ecu 

■* —4 •»< C 

CM 00 V 

to • C • b 

•4 CM U U b 

•« CCS 

« • U 

b b U 

b b O 

3 3 

O " 

U U to 

O O O 

00 c t ^ 

^ O •J a: 

« a> o 

c c » 

«e O O H 



CO «o 









CO w 



ca to 



CD 






CO 



CO 

CO 



CD 



00 c« 



o 



•^00 






in 
«o o 



(D 

to <o 









♦ t- 




lO 


CO 


♦ ^ 


C4 lO 




«D a> 


NIO 


C4 CD 


• 




• 


• 


• 


00 




to 


lO 


to 


CM 




«H 






♦ o> 


<D 


to 


to 


* a> 


'•* C4 


«^ to 


C« CD 


eo to 


too 



to 

I to 



eo 
to 



»-i o 



o> 



♦ to 

CO (O 



« M 



^ »~ "< CO 
C4 • t-l . 



«o 

lO CO 



eo 



o 

CO CO 



■♦ o 



to 
O eo 



lO 

00 



eo 

CO 



eo 
00 



00 



o 
o 



CO o 

CO . CO o 

CO 

«o o 

o 



eo 

CO 



C4 eo 



CM ^ 

CD 



00 
CM 



CO 
CO 






u 

« O) O 

o> ^ • • to 

CM *» 

X x « 

U U %i 

• eo o O 

Ok O to tB C 

CM • e « -^ 

00 ^ <*' 

« « o 

O O B 

c e « 

o 

O O b 

a m « 

« « CD 

m n X 

■^ •>< b 

o 

CO CO to 

« 

X x^ 

b b « 

o o u 

to to 
« « e 

« « 
o u ■ 

V 

e e V 

e> .^ ••« e 

o »- « 

^> • « O b 

M « U U b 

r> c c 3 

V V u 

b b U 

b b O 

3 9 

U or 

u u to 
o o o 



00 
CM 



CO 

eo 






eo 



u 



11^ 



to 
o 



13 



« o o 

CCS 
OOH 



o 
o 



o 
o 



o 

C4 O 



o 



e 



o 

lO 



e 

M O 



O 

to 



to 

•4 c^ 



n 
m 



CD «D 



CO 






o 


o 


o 


o 


o 


o 


o 


o 


o 


t«. 


»» 


« 


o 


-» o 


c* o 


^ o 


•H O 


«D O 


CO o 


MO 


^o 


NO 


o» « 


l« CO 


• 


• 


• 


• 


• 


• 


• 


• 


• 


• 


«4 


^ • 


o 


o 


o 


o 


o 


o 


o 


o 


o 


«D 


o 


« 


o 


o 


to 


lO 


o 


o 


10 


o 


o 


«o 


t» 


CD 



o 

Id 



td 



117 






o 



^" 


T 


N 


«4 


o 


to 


N 


lO 


to 


U> (D 


CO «o 


^ OJ 


r> o 


«o 


M 00 


•H ca 


• 


• 


• 


•^ • 


* 


• 


• 


• 


« 


O 


CO 


€>» 


O 


•V 


«4 


<o 



CO 


«4 


lO 


o 


eo 


e>» ^ 


CO eo 


CD oa 


• 


• 


« • 


ej • 


c« 


to 


la 


to 



00 

C4 ir> 



O CD 



o 


c» 


o> 


^ 


o> 


CO 


CO 


»» 


o 


eo 


« o 


o •» 


to t^ 


o»o 


to 00 


to CO 


o> C- 


^ CO 


M O 


<^ CO 


^4 • 


e>» • 


M • 


ei • 


ID • 


e>» • 


to • 


• 


ff4 • 


^4 • 


o 


•-4 


CO 


CO 


•4 


CO 


o> 


CO 


to 


CO 


lO 


««- 


ID 


CO 


ID 


eo 


t- 




e- 


CO 



•^ eo 

I '^ o o 

to C4 eo 



o 

lO o 
CO • 
^ CO 

CO 



to 

•-< eo 



coeo 



to 
i eo 



eo 
CO eo 



10 



CO 

•-4 O (4 1 



eo 


e 


oo 


at eo 


CQ • 


^* • 


CO 


« 



CO 


CO 


to 


o 


CO lA 


^ eo 


c» -« 


c* o 


C4 • 


v4 ■ 


^4 


• 


o 


lO 


CO 


to 


to 


^ 


^ 


CM 



a> 


00 


t- 


CD 


o 


to 


^ 


CO 


eo 


et 


o 


cj o 


•♦o 


e« CO 


-«eo 


a> o 


CO e» 


>4 CO 


at 00 


o o 


oco 


«» ^ 


«i4 • 


CO • 


• 


•M • 


0> • 


• 


e«» • 


C»I • 


CO • 


•^ • 


eo • 


CO 


M 


CO 


■» 


o 


00 


CO 


CM 


«« 


CO f 


CM to 


C4 


CO 




"< 


a> 


^4 


CO 


eo 


eo 


^ 


to 



00 

CO 



CD 
C4 



00 
CM 



CO 

o 



o 

CO 



CO 

eo 



to 
eo 



o 


o 


to 


o 


CD 


to 



w 



u 



u 



1T8 






(O 



« 









o> 



eo 

C4 • 



M O 



o 


w 


in 


v4 


lO 


t- 


O 


to 


t" 


eo 


<o 






«>» 


10 


CO 


ex 


e- p» 


(D 0> 


lo r> 


U> -4 


N t^ 


o «-• 


o» P- 


•» o> 


(oe> 


•M eo 


•^ t» 






» ex 


-< eo 


v« aM 


CO la 


• 


• 


• 


• 


^4 • 


e>» • 


e>» • 


ex • 


ex • 


• 


• 






• 


• 


<» • 


eo • 


e« 


en 


e>« 


eo 


eo 


la 


^ 


^ 


ex 










ea 




^ eo 


— < ex 


to 


c« 


^ 


^ 


eo 


t- 


^ 


V 


<D 


ex 


eo 




lO 


eo 


lO 


O 


o 


in eo 


t- o 


a> o> 


eo t~ 


<8 n 


(O o> 


o> t~ 


eo t« 


ex-* 


•<f o> 


O to 


o> CO 


eo ^ 


t» ex 


eo -4 


— lO 


at • 


(O - 


(C • 


^ • 


c* • 


9^ « 


«0 • 


e- • 


^ • 


o> • 


T • 


eo 


• 


^ • 


lO • 


^ • 


to . 


v4 


CD 


t- 


«o 


-« o> 


^ o> 


ex eo 


^ to 


"^O 


o 


O 




«o 


^ 


t<- 


CO f 


t«» ex 


« 


CO 


r» 


C4 


eo 


ex 


eo 


eo 


u> 


eo 


eo 




ex 


«M 


CO 


«-« eo 


^^ CO 


« 










(O 


















lO 


t- 


*- 


-< « 










«H N 


















-< CO 


coo 


o» -N 



I 



eo 




CO 




eo 


eo 


^ 


CO ^ 


ex 


M 


^ 


CO 


00 eo 


00 ex 


CO • 


t» 


• 


^^ 


• 


m>4 • 


V • 


to 




o> 




t- 


•^ CO 


-* 00 


«o 




CO 




ex 


CO 


eo 


eo 




to 




o> 


CO 


o> 


ex o 


eo CO 


a> 


to 


lO lO 


to ex 



Oi O Oi t" 

•-« o •HO '••ex t«- o 

co« ex» o» eo« 

M CO o4 to ex CO ^ to 

ex ex ex ^ 



eo V 

ex eo «» ex 

eo . CO • 

*" eo 



CO 

CO 00 



to 
ex o> 



eo ex 



eo 
O) o 



~4 to 



O — 
a> • 



to eo 
ex 



o 

CD CD 



le ex 



eo ^ 
^ eo -« to 



CO 
C4 CO 



eo ««• 
eo o -^ to 



'9' ex 
-< eo 



ex 

' CO 



to 



ex •< 
ex •* 



^ CD 

** • 

-* ex 



eo 
eo CO 



^4 


CO 


«o 


t- 


^* 


CO 


^ 


e> 


lO 


^ 


ea 


o 


to 


t- 


eo 


eo 


OCO 


"W CO 


CO -* 


O ex 


t- 00 


e~ o 


eo o 


to CO 


f- eo 


o«» 


* eo 


eo eo 


eo t» 


eo to 


a> ex 


ex to 


O) • 


•V • 


eo • 


CO • 


to • 


o> • 


t- . 


o • 


CO • 


CO ♦ 


cx • 


CO • 


eo • 


eo • 


cx • 


to • 


o> 


eo 


eo 


e- 


r» 


to 


'•* eo 


«-« vM 


«M 1M 


a> 


00 


ex 


eo 


•14 


O CO 


o a> 


ex 


ex 


— • 


CO 


•H 


ex 


ex 


c» 


e« 


^* 


«M 


ex 


•H 


ex 


-* ex 


«H »4 



•V 


«o 


CM 


o 


00 


00 


eo 


— 


•— 



o 
cx 
n 



00 
eo 



CO 
00 



o 

CO 



eo 
to 



0> 
ex 



00 

to 



Cd 



119 



o 

CO o 



«D 



in C4 






M 



o> 



o 



to 



o 

lO 

lO 






(4 

e> o 



O 
2 





































« 




































u 


o 


o 


o 


o 


o 


CJ 


p» 


o 


o 


O 


O 


o 


o 


o 


CO 




W4 


e 


to o 


CM O 


« o 


o> o 


lO o 


la ro 


W> P> 


N O 


(D O 


e* o 


« o 


o> o 


lO o 


to o 


to » 


^ p> 


o 


• 


• 


• 


• 


• 


• 


• 




«4 • 


V4 • 




• 


• 


• 


00 • 


'^ 


• 


(4 


o 


o 


o 


o 


o 


CO 


« 


o 


O 


o 


o 


o 


o 


o 


m 


V4 


e<» 


h 


o 


o 


o 


o 


o 


eo 


eo 


o 


eo 


o 


o 


o 


o 


o 


e» 




e> 


3 


«4 


«4 




w^ 


• w^ 


























U 
U 

o 
o 


lA 


N 


ri 


Oi 


•n 


(D 


to 


OS 


o 


C4 

^4 


« 


o> 


to 


lO 




to 




o: 

: 
V 

e 
o 



u 



120 



o 
• 


• 






o 
00 n 








at 



CM 



00 'V u» o> 



(D 
CM 



CM »H 







^4 




CM 






o 




o 




(D 


CO 


CO 




CO 


00 




tn to 


-^ 


»» 




V 


CO 


^4 


o 


-* o 


CM CO 


CM O 


(O 


PO 


«^ c~ 






* 




• 






• 




• 




• 


• 


• 


CM 


• 


CM • 






•M 










CM 




v4 




W) 


CM 


CM 




•H 




«0 




to 




t* 


CD 




CM 




o 




«o 


(O 


00 




t- 


lO 


CM CM 


<0 CO 


c» o 


O) -V 


OJ o> 


o> o 


00 


v4 


O CM 


0> CM 


in 


^" 


r~ lA 


«^ * 


CM 


• 




• 


^4 « 


vi 


• 




• 




* 


c« • 


• 


(D 


• 


in • 


o 




00 




lO 


(D 




o 




o> 




o 


CO 


a> 


wm 


00 


CM o> 
















«i4 








'" 


CM 










10 




CM 






* 














(D 






v4 


CM 


-H 00 


^^ 


CO 






•^CO 














f-« ^4 




■^ 


CM 


CM 00 


• 




• 




















«-4 






• 


CM . 


t- 




o 




CD 


CO 




a> 




o 




CM 


CM 


•V 




00 


t» 


T at 


CM 


to 


v4 


■V 


00 t- 


00 o 


CM O 


»4 


o> 


CO •< 


-< CO 


00 


CO 


CO t- 


«4 • 


^ 


• 


CO 


• 


lO • 


CM 


• 


•H 


• 


«4 


• 


«4 • 


•i< • 


lO 


• 


mm • 


^^ 




CO 




N 


o> 




(D 




CM 




CO 


in 


«M 


CO 


eo 


(O CM 


«4 




•-« 




CM 


«4 




«H 




w^ 




«i4 


vM 


«4 




<^ 


CM 



in «o 

^4 in **4 ^4 



in 

— CO 



CM 

^H CO 



CD C- 



CO ■«• 



a> in 
I in «4 CO 



C4 



CO C« 

»T CO "H in 



o 
-^ o 



CO 

CM in 



m* CM 



CO 


o 


<-<o 


at o 


• 


CO • 


^•1 


CM 



in CO 

CO • 



ta n ■m* a> ^•^> '4in^ooe>i-«in«M 

inc~ CMco <o<T o>CM incM incM incD>CM« co^r C4o> coo otco moo in^* cd^- 

• • • • • • • •«b4«^4« • • • aQO* 

eacM^^ COT •-<»<in«oco»^ininT 



« o 

CM -4 



CM 
CO 



« O CM O 

in -< CO c<- "HO 



to «H «4 



«a> oo«c> in 00 t« in t~«' 

o^ «DO» •-•^ 'Tco o>ao mcM «-it- «-io> -<. 

^. . . . . .eM«<-i>eM 

incD^cD in^p (Dt>i 



e- o 

0> «4 CM o 



in CM 

CM • 



O) CM * T 

^ ^ T -« CD ^ 

o • 

CO T -^ in 



in 


«» 


o> 


»~ 


T 


«» 


CM 


^ -< 


CM 00 


•^ CM 


N ID 


-" o» 


00 CO 


in o> 


00 • 


T • 


T • 


T • 


00 • 


r- • 


o • 


(D 


00 


o 


CO 


CM 


CO 


CM in 


^^ 


"T 


<o 


(D 


in 


<o 


CO 



to 


CM 


o 


«o 


^ 


«4 


»» 


lO V 


00 CO 


— o 


in o> 


CM 00 


mo 


•4 m 


CO • 


o> • 


t* - 


»r • 


^^ • 


CD • 


T • 


^«D 


CO 


«i4 


o 


00 


e» 


— 00 


•n 


in 


«>- 


•n 


^ 


CD 


-4 in 



CM 

00 



CD 

00 



00 

to 



CO 

<o 



CO 

in 



00 


T 


CO 


o> 


>4 


CM 



o 
o 



00 



m 




T CO 


• 


^r • 


t" 


-« m 






>» 




• u 


» o 


e- o 


OD CO 


. . to 


• 


t-c~ X © 




U *i 




X >> o « 




U U ttt o 


» t- 


o o o 


m CM 


wte-u e 


mm • 


V V tf-** 


— « T 


^ 4J u 




« « o 




t> u e m 




v4 ^ 




e e « 




»4 94 O 




to « 




O O-^ fci 




n « e « 




^H (i^ 


- 


« « oco 




Im 




n a « X 




•^ «4 b 




«o O 




coco M 




>. « 




X X (< 4J 




b l< o « 




o o «o u 




«o «o o 




« V A> C 




4J -kJ « •" 




« « o 




o o « 




e V 




C C-H o 


CO 


^>« c 


CD 00 


n « 


e~ • 


V W V b 


T T 


u U U b 


mm lA 


C C C 3 




W V V u 




u u u u 




U U Li O 




3 3 3 




V V V- 




U U U «0 




o o o o 


CM 




o> 


I s s "O 


(O 


< t)-^ o: 


CM 


: : : = 




V V X o 




c c-« > 


«0 


O O to H 



o 



121 



— o 
ca 



eo 
ca o 



o 



o> 


o 


CO 


>4 


«o 


^ 






eo 


o 


N O 


p» eo 


«4 «« 


t- ^r 


-<«o 






-« t» 


• 


• 


• 


«4 • 


• 


• 






• 


o> 


«0 


N 


(O 


n 


«i« 






N 


o 












•-« 


in 


eo 





o o 

n • 

to 


lO 

e«» o 


eo 


0> 00 





w^ at <-• (o 



ea ^- 



u> o 



eo 
to eo 



CO 



o 


o> 


o 


c« 


o 


« 


0> 


to 


eo eo 


u> o 


o> eo 


e» o 


ca M 


r> eo 


• 


• 


• 


• 


N • 


o • 


• 


««• 


eo 


o 


»4 


M 


M (O 


eo 


eo 


eo 


o 


eo 


e> 


o> 


to 



M 


eo 


•o 


e> 


o 


«o 


•M 


«o 


eo o 


eo lA 


ca eo 


(O o 


e> eo 


«^ ^4 


• 


to • 


c» • 


« . 


CO • 


CO • 


lO • 


* 


lO 


«o 


eo 


o 


ca 


^ -< 


eo 


o> 


e> 


eo 


o 


o> 


e» 



Ok 

eo eo 






CI 



o 



ca 

«o 



Ci 



CO 









Ed 



122 



n 



n 



a> 

I o 



« o « o> 



o 






CO 
•0 



« o> 
eo CO 



C9 






I o 



« o 
CO o 



o 









« ea 



* o> 

eo 00 



eo 

































XX 
































%* u 


























; . * ,', 






o o 






















i 


-.<" 






:;:.; ' 


tete 
• « 

ss 

e e 






















~ 1 




\ -, 






o e 










o> 






eo 




e> 








M 


eo 


n ■ 










>^ o 






^ eo 




f-4 CD 








« eo 


CO eo 


•^^4 










• 


•* " 








• 










• 


c « 










o> 






lO 




«- 








eo 




• • 

u u 
« « 

•O<0 

XX 
b b 
O O 

« » 


o 


e* 


e 


o 


e<* 


o 


o 


M 


O 


CO 


r- 


o 


o 


to 


e- 


u u 


o> o 


C4 «D 


« o 


-« o 


at eo 


in o 


-« o 


«o ^ 


eo o 


•^ to 


roto 


lO o 


e- o 


'TOO 


^ ^ 




»< • 


• 


• 


• 


• 


CM • 


• 


f^ • 


V4 • 


«i4 • 


«4 • 


• 


• 


CO • 


o • 


c e 


o 


(O 


o 


o 


"^ 


o 


o 


^ 


o 


V 


(D 


o 


o 


«4 «4 


CO a> 


••4 ••4 


o 


(O 


o 


o 


eo 


o 


o 


a> 


eo 


eo 


eo 


o 


o 


e> 


eo 




-« 












«4 










«M 


«« 




• 


« » 

V V 


o> 


r» 


t^ 


»* 


^4 


in 


^4 


e» 


lA 


« 


lO 


•A 


e- 


lO 


CO 


3 3 


«-4 








^M 


ex 




«>4 


«-« 


•^ 


^4 






>4 


eo 

CO 


o o 



w 

ce 

td 



< < 



123 



OS 



The foregoing tabulations give totals, for each Brown Corpus 
genre of our sample, of the incidence of each lower-sign sequence 
in all 18 of its possible manifestations. These totals vary in 
interesting ways by genre, and the category percentage is seen to 
vary considerably also. So, for (be) category 1 varies from 5.41 
to 33.33%, and the sparser category 5 varies from 1 to 13%, while 
category 12 hovers between 1 and 10%; yet the more populous 
category 14 shows a more stable range of 3.51 to 20.19%. However, 
it will be seen by closer inspection that stability does not go 
simply together with frequency. The differences in these ranges of 
variation must have to do with contrasts elicited by the rules and 
constraints of genre and with the content of English linguistic 
form. Remember that the structure and rules of braille itself 
govern the very appearance of any one of categories 1-18. 

This type of variation is rather less rich for (to) and (in) . 
In the sparser items such as (by) , (enough) , (Into) , and (were) we 
find far less in this dimension to work with. These superficial 
remarks are not intended to deal adequately in any way with the 
aspect of genre, category, and linguistic formational variation, 
but only to call attention to this fascinating textual parameter 
and to its clear manifestation in our modest data. What is needed 
is the collection and analysis of a much larger body of such 
contrastive data, and a long period of study which will surely grow 
out of such a probe. 



124 



Patterns of Incidence 

If we now disregard this category variation and concentrate on 
the totals of category incidence for each lower-sign sequence, let 
us get a better visual grasp of the incidence of these categories 
in our sample corpus. To do this we will translate the sequence of 
category totals given for each lower-sign sequence into a template 
of line graphs. The problem of estimating the relations in 
behavior of these lower-sign sequences then becomes a simple one of 
visual template fitting, which seems to suffice for our immediate 
purpose. 



125 



BE 



CATEGORIEi 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 



I • 



100 



90 



70 



E— ■ 



O 






50 



30 



—\ 



I 1 



I r 



I 1 



; ± 



I !■ 



r ! 



! -+ 



! V 



L 1, 



L. 



1 



4 



"~l" 



I r 



1 1 



1 — r 



t — !• 






! 4 



■! + 



! 1 



r~4 



~i' 



I — h 



I V 




-4-+ 



i 1 



L. 



I r 



I -t 



1- 



1 1 



1 \ 



I ^ 



! y 



i 1 



! V 



J 1 



I 



f-H 



I A- 



i 1 



7 154 



108 29 U7 165 247 



OCCURRENC] 



126 



BY 



CATESORIE! 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 



100 



90 



U32 



E— ■ 



P4 

o 



pqi 



60 



50 



40 



30 



Ll 



Ll- 



10 







h- 



i J 



J 1 



I 



4- — ! 



i 4 



1 



J I 



1 L 



L I 



•4 



r — I 



T ! 



1 1 



-h 



I r 



r ! 



t i 



J 1 



J 1 



I. 



i 1 






I-- 



! a. 



1 » 



.1 



-f 




~1 



1 



— I 1 — 



T 1 



J I 



1 ! 



: -h 



L 



J I 



I JL 



I_. 



I I 



I 1 



'1 



T 1 



I 1 



127 



ENOUGH 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 



100 



E-=^ 



O 



pqi 



50 



30 



10 



i 


It'' 

i. _! 1 1 


1 1 j in T^i 1^ iHT^i 
1--+ u-i LJ L_j i._j _! i._i 


i- _4 ' ' 


i__i i__| '_! j _j 4. 1 1 _!_ 1 


1 ^ T -| 


170^1 1 1 ' i 1 1 1 i ' i ! 1 
i — r~i — ^— I — r— ] — \- — ! — t — ' — [ — I — i 

L — j — i_ — 1 — 1. — { — ; — j — I — 1 — 1 — 1 — 1 — 

|-l— 1 { ■\ 1 -}— J- 1 {- 1 r 1 f 


■ 1 t "~ T 

- — ! — r — 1 — i 
— 1 — i — T — 1 
— [ — 1 — 1 — 1 

L J. U -! 


1 



-i--- 






1 I "^ ^ 
J L J _ J 


'11 i/m' ! 1 ' I 1 ' li 


II 1 ^ 

L _L L ' ! 


( \\ 1 / 1 \' r ' ■ ■ [■ 1 1 1 ■"■T ' I 

' — \{ — T" "•" 1"' — ] — i — "^ — ' — "{ — r — i — ^ — 
' 1 / -I \ I I 1 1 1 i i 1 


_ J 



19 



OCCURRENCl 



128 



HIS 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



100 L_J 



90 



E-« 



o 



p!^ 



80 



70 



60 



50 



30 



20 



10 








129 



IN 



1 2 3 4 5 6 7 8 9 10 U 12 13 14 15 16 17 



100 



90 



70 



. E-^ 


S 


pqi 


o 


. o^ 


1 pqi 



50 



40 



30 



20 



10 



I 



L_. 



\ 1 



1 1 



_J 



I L 



H 



T 1 



1 ! 



T 1 



I r 






J 1 



J 1 



I 






I 



4- 1 



\ 1 



; 1 1 + 



i , 




~1 



~l 



T !■ 



r ! 



T 1 



J 1 



,t J. I 



I 



4. ,. 



I 



\ 1 



"T 



( 1 



T 1 



1029 



14 86 1447 3 1643 141 



66 



130 



INTO 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 



IP4 
O 



P4 



i 1 1 1 

100 L ! i- -i 


II'! 


1 1 1 1 - 1 1 1 
i- — 1 JL ' i. _i J. _j ' 


1 


193.48; 1 1 1 1 1 1 

90 I 4 _ L i L J {._ J 


1 iH ^1 i^ 1 

4-— 1 -L— ! i-— I i-— 1 ' 


1 


II iT ~i 
I 1 1 1 1 ' j 1 

80 Lj.-|— [-j--i--r--i 

70 jLL.]___L_^___L_j___ 

i i i i M i i 

60 L. .__{ — L__ J — [ — j — |__J 
50 L. _4 — L__i — ^__4 — j. — 
40 i . i J [J i 4- . 


1 1^ I 1 1 

■ i t "• "^1 t "t" 1 i 
— 1 — {_ — } — f — 1 — I. — 1 — I — 

— t — i — \ — \ — 1 — \ — 1 — h""" 

— L — i — {_ — 1 — I — 1 — L — 1 — 
1 { 1 1 j 1 j 

J 1 1 j { II I 


" — 1 
1 

i 

1 


30 [ . 1 1 L 4 j. 4 


__i j_4 r4 ^-J 4-- 


r 1 


20 i J I J 1 ! i 


1 j ! ! 1 1 1 - 1 




r 1 1 -j- ■ r -t - — 

10 I A Is i LA . 


1 L_i L_l u-l L_i 


r 1 


lLL 


1 
1 1 1 

» 1 1 I 

trts-J 1 


T ii r^ iH It 

1 1 1 1 1 1 1 I 1 3^ 
1 1 1 1 1 1 1 1 l^ 


r "i 

y 



86 2 1 



131 



TO 



CATEGORIES 



1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



100 L_J. 



i 1 



I 1 



70 



J I 



o 



50 



PL3 






— (- ! 



30 



20 



10 



-i 4 



I I 



1 1 



1 1 



1 1 



I r 



•r- 



— -t- 



r i 



I y 



J 1 



J 1 



I 



X 1. 



I 



4. 1 



J 1 



I I 



J 1 




I I 



1 1 






1 1 



1 ! 



J I, 



I 



4. 1 



J I 



J. 1 



^~H 



I 4 



T 1 



I 1 



106 25 86 39 6 



OCCURRENCE 



132 



^;^S CATEGORIES 

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 



100 L..I 



9U1 



70 



i-i — 



PL3 

o 

PC] 
Pi 



50 



30 



20 



10 



I T 



r~T 



L 



X 



I + 



1- 1 



! -4- 



L 



■i 4- 



i 1 



\ J. 



I r 



I T 



I -r 



i 1 



I 1- 



I h 



ibi 



I — I I I i i 

451 






I 



! i 



t-- 



! -t 



I 1 



I 1- 



I 1. 



I 1 



I 



I r 



I 1 



I 1 



I \ 



1 1 i i 



I 1 



I 1- 



! ^ 



I 1 



! r 



— t 



1- 



\ \ 



I h 



! 1 



6.^6 




I i 



I 1 



I 



._j 



* ■ 



9 30 



OCCURRENCE 



133 



WERE 



CATEGORIE 



1 2 3 4 5 6 7 8 9 10 U 12 13 14 15 16 17 



O 



It 1 1 t * t 1 1 

100 L_|_i_J__4--|-4--J---i 

I^LSSJ 1 1 1 > 1 1 1 

90 1_ 4 _L J _L 4 i J 1 


— 1 
■ 1 1 ■ t 


1 
1 
1 

1 
1 

1 

1 


1 
1 
1 


1 
i 


,^ 1 1 1 1 1 1 1 1 
80 1 — }— 1 — l— 1 — f — i — 1 — 1 

70 I. — 1 — 1 — L — J — { — } — j_ — 

60 L . 4 i. J .J [J \ 




— -j— j— 1 
— 1 — |. — 


I 1 i T 1 

— 1 — h~T — \ — -j 

— [— i — f~i — \ 

, . 1 i._ 1 1 


OU LJ i L J. ' J _ 1 -I -i 




r ■ 1 1 


11 [ 1 




. 40 I J-i. _J _i. ! 1- J JL 




- r ~i ^ 


u- JL-, L_ 4. L 


~i 


vU [■■ l 1 ' ' ' { 1 


_ 


i_ * i_ 


r 1 I I f^ T 
! L J 1 J i 


20 |._li 1 J_ i J_ >L ' 


[ 


}- — I- r 


! _ L -L 1 I ! 


10 L 1 L 4 I i L ' - 


L _ 


J. J. ... , , . .. 


ill 1 ^ ' 
! J _L J J 1 


r f 1 1" j [ j r 

1 1 ! iP 1 1 1 1 


I— 


I— i — I— 


t 1 ite 1 if 

1 yt^ 


1 
X 1 



124 



8 8 



OCCURRENCE 



134 



In a visual inspection of these displays it may be useful to 
recall the frequency groupings: 

frequency rough congruence 

rarest (enough) , (were) (into) 

rare (by) 

median (was), (his) 

most frequent (be), (to), (in) 

This may provide a useful association for purposes of teaching 
and mastery of these various manifestations and functions of lower- 
sign sequences; but the specifics must be discussed separately. 
Braille and Inkprint 

We now proceed to study the incidence of lower-sign braille 
units in relation to print letter sequences. 

Let us view the data which have been presented in detail in a 
more summarizing fashion by displaying the braille disposition of 
the print letter sequences corresponding to our lower-sign braille 
units . 

First, some notes recalling the representation in braille of 
print occurrences of the letter sequences which correspond to the 
lower-sign words. The table indicates the ways in which the letter 
series in the text scunples are transcribed into braille. 
The text sources, selections from the Brown Corpus and an issue of 
the i^eader's Digest, are specified by "BC" and "RD." ' 

In the first numerical column there is a count of the 



135 



occurrences of the letter sequence. 

The second column gives the count of the lower sign braille 
contractions . 

Sometimes the letter sequences under consideration occur 
within a longer sequence which has its own braille contraction, as 
(be) in (before) , in in (blind) , or to in (together) . The count of 
such inclusive contractions is given in the third column, and it 
should be noted that the count of this type of sequence is not 
included in the second column. 

The fourth column of figures shows the number of times the 
sequence is represented by being partly contracted. This figure is 
independent of the two prior colximns. 

In some cases the first or last letter of a series is combined 
with an adjacent letter or group of letters to form a contraction 
which overlaps the lower sign series. Examples are the overlap of: 
by by (bb) in lo(bb)y, in by (ness) in busi(ness) , to by (st) and 
(one) in (st) (one) , and was by (sh) in wa(sh) . The count of this 
type of occurrence is in column five. 

The last column shows how many times the letter sequence is 
found uncontracted, that is, represented letter for letter in 
braille. 



136 



print 






lower 


inclu- 




over- 


letter 






sign 


sive 


partial 


lapping 


se- 






con- 


con- 


con- 


con- 


quence 






traction 


traction 


traction 


traction 


BE 


BC 


1126 


362 
32.15 


154 
13.68 




247 
21.94 




RD 


1184 


362 
30.57 


169 
14.27 




287 
24.24 


BY 


BC 
RD 


287 
361 


242 
84.32 

270 
74.79 






2 

.70 


ENOUGH 


BC 


27 


19 
70.37 




8 
29.63 






RD 


18 


15 
83.33 


.- 


3 
16.67 




HIS 


BC 


650 


310 
47.69 


280 
43.08 




38 
5.85 




RD 


500 


287 
57.40 


165 
33.00 




26 
5.20 


IN 


BC 


4429 


2813 
63.51 


1533 
34.61 




66 
1.49 




RD 


5387 


2939 
54.56 


2393 
44.42 




26 
.48 


INTO 


BC 


92 


86 
93.48 




6 
6.52 






RD 


156 


144 
92.31 




12 
7.69 


1* 
.64 


TO 


BC 


1948 


1141 
58.57 


125 
6.42 




191 
9.80 




RD 


2692 


1476 
54.83 


179 
6.65 




278 
10.33 


WAS 


BC 


495 


451 
91. 11 






30 
6.06 




RD 


814 


748 
91.89 






33 
4.05 


WERE 


BC 


135 


124 
91,85 




11 
8.15 


8* 
5.93 




RD 


228 


204 
89.47 




24 
10.53 


22* 
9.65 



no 
con- 
traction 

363 
32.24 

366 
30.91 

43 
14.98 

91 
25.21 



22 

3.38 
22 

4.40 

17 

.38 

29 

.54 



491 
25.21 

759 
28.19 

14 
2.83 

33 
4.05 



* This count is included in the "partial contraction" count 

on the same line. Braille representation in these occurrences 
involves both contraction overlap and partial contraction. 

137 



In partly analogous fashion we now inspect the incidence of 
"whole words" and these lower-sign braille units. The braille 
disposition of the lower-sign words occurring as whole words in the 
print text is shown in the following table. To render the results 
comparable for purposes of this and the following tabulations the 
Brown Corpus interpretation of "whole word" has been accepted 
throughout . 



138 



Representation by Lower-Sign Units and Other Braille Units 
of Lower-Sign Letter Sequences Occurring In Print as Whole Words 



print 
whole 
word 

BE 



partly 



BY 



ENOUGH 



HIS 



IN 



INTO 



TO 



WAS 



WERE 



Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 

Brown Corpus 
Reader's Digest 





con- 
tracted 


con- 
tracted 


uncon- 
tracted 


306 
224 


225 
73.53 

152 
67.86 






81 
26.47 

72 
32.14 


270 
306 


242 
89.63 

270 
88.24 


p 


d;: ■' ; 


28 
10.37 

36 
11.76 


27 
18 


19 
70.37 

15 
83.33 


29. 
16. 


8 

63 
3 
67 




331 
308 


310 
93.66 

287 
93.18 






21 
6.34 

21 
6.82 


1043 
1070 


1029 
98.66 
1052 
98.32 






14 
1.34 

18 
1.68 


89 
151 


86 
96.63 

144 
95.36 


3 
4 


3 
.37 

7 
.64 




1264 
1597 


1141 
90.27 
1476 
92.42 




.. 


123 
9.73 
121 
7.58 


456 
762 


451 
98.90 

748 
98. 16 






5 

1.10 

14 
1.84 


127 
206 


124 
97.64 

204 
99.03 


2 


3 
.36 

2 
.97 





139 



Relative Frequency 

Now that we have inspected in detail some important relations 
between the occurrence and non-occurrence of these lower-sign 
braille units in relation to relevant print letter sequences, it 
might be of interest in retrospect to summarize the frequencies of 
these units in the text studied. 

The occurrences of the lower-sign, one-space (or, in the case 
of (Into) f two-space) contractions are ranked by frequency in the 
following table. 

Lower-Sign Word Contractions 





Brown 


Corpus 


Reader ' s 




Sample 


Digest 


IN 


2813 


50.7 


2939 


45.6 


TO 


1141 


20.6 


1476 


22.9 


WAS 


451 


8.1 


748 


11.6 


BE 


362 


6.5 


362 


5.6 


HIS 


310 


5.6 


287 


4.5 


BY 


242 


4.4 


270 


4.2 


WERE 


124 


2.2 


204 


3.2 


INTO 


86 


1.6 


144 


2.2 


ENOUGH 


19 
5548 


0.3 


15 

6445 


0.2 



Seunple Consistency 

At the sctme time we have seen in the foregoing table the 
gratifying agreement in overall incidence of these nine braille 
shapes between the Brown Corpus Scimple and the Scunple from the 



140 



Reader^ s Digest. Comment on the small deviations between these may 
be reserved for another occasion. 

As a test of consistency in the translation, identification, 
and categorized collection of data in this study, we offer the 
following comparison of the constituency of our data for lower-sign 
sequences occurring as whole words with that published for the 
Brown Corpus (KuCera & Francis, 1967). 

To render the results comparable for purposes of this 
tabulation the Brown Corpus interpretation of "whole word" has been 
accepted. For the same reason the data from our scimple and from 
the i^eader's Digest result in every case from summing columns 1 
through 7 of the 18 displayed in our categorization. 





Brown Corpus % 


Brown Corpus Sample 


Reader ' s 


Digest 


TO 


26149 


32.1 


1264 


32.3 


1597 


34.4 


IN 


21341 


26.2 


1043 


26.7 


1070 


23.1 


WAS 


9816 


12.0 


456 


11.7 


762 


16.4 


HIS 


6997 


8.6 


331 


8.5 


308 


6.6 


BE 


6377 


7.8 


306 


7.8 


224 


4.8 


BY 


5305 


6.5 


270 


6.9 


306 


6.6 


WERE 


3284 


4.0 


127 


3.2 


206 


4.4 


INTO 


1791 


2.2 


89 


2.3 


151 


3.3 


ENOUGH 


430 


0.5 


27 


0.7 


18 


0.4 


TOTAL 


81490 




3913 




4642 





141 



It will be seen that except for the three least frequent the 
deviation among these percentages is very small. The greater 
deviation for the J^eader's Digest material is easily understood as 
reflecting the different and less systematic genre representation 
among its texts . 

It might be of interest to readers to compare the decrement in 
incidence of these lower- sign words in the Brown Corpus with that 
displayed in the consolidated Lorge-Thorndike counts (Thorndike and 
Lorge, 1944) for these words individually. 



"" ; , 


1 


2 


3 


4 
Rounded 




Brown 


Lorge 


L-Th 


Average 




Corpus 


Magazine 


Semantic 


(2 & 3) 


TO 


26149 


115358 


?* 


115M 


IN 


21341 


75253 


96674 


86M 


WAS 


9816 


58732 


42552 


51M 


HIS 


6997 


30748 


32140 


31M 


BE 


6377 


19645 


?* 


20M 


BY 


5305 


11454 


29130 


20M 


WERE 


3284 


15082 


16340 


16M 


INTO 


1791 


9231 


7016 


8M 


ENOUGH 


430 


2113 


892 


2M 



*This means that data were not available from this count. In the 
Preface to Thorndike and Lorge (1944) we read that "the recorders 
of the semantic count . . . did not separate different forms of 
words like be , come , and do , . . " . 



142 



All these words are among the first 500 words and except for 
enough are registered as totalling 800 times or more in the 
Thorndike 1931 data and 1000 times or more in the count of 120 
juvenile books. In the case of enough the count was 800 or more 
for the 1931 data, but only estimated at over 1000 for the juvenile 
books. All these words rank at 100 or over per million. 

The congruence in the decrement profile for these words in 
these corpus counts is reassuring. -' . i 
Some Further Conclusions 

The question may be asked how salient or important the lower- 
sign words or braille units are in the braille code. Such a 
question deserves a carefully considered answer, but there are 
different ways of answering such a question. One may, for exsimple, 
consider such a question from the point of view of the overall 
efficiency of the code, or as a function leading to the learning 
and mastery of other elements or aspects of the code. Such 
considerations need not take us up and detain us now, but can 
certainly be usefully addressed at a later time. 

For the present we shall content ourselves with a simple 
quantitative observation. Our Brown Corpus sample comprises 50,000 
words of l^erican English of which 3,913 are whole words of the 
lower-sign set discussed in this paper? this means that the total 
lower-sign words comprise 7.8% of the sample text, and that ca. 
.57% are lower-sign words which are uncontracted (i.e, not 
represented by lower-sign braille units for contextual reasons). 
These figures compare favorably with those for the full Brown 



143 



Corpus, which comprises 1,000,000 words of which 81,490 are whole 
words of our lower-sign set, or 8% of the total text. 

We thus arrive at a quantitative statement of the proportional 
occurrence of the objects of our study. Out of 50,000 words 7.25% 
(= 7.8% less .57%), or 3,627 words, are represented by lower-sign 
braille units. Out of 180,000 (actually 179,429) braille units 
, comprising our sample text 3% (5,548) are lower-sign braille units. 
In other words, one in every 32 braille units of text will be a 
lower sign; one may expect to encounter on the average a lower-sign 
,, braille unit in almost every line of braille. 

.-jj^ . Given the textual frequency which we have observed for these 

- lower-sign braille units, and consequently for the lower-sign 

shapes, as a group, it is clear that for any learner of braille the 

mastery of these braille units and shapes forms a considerable 

portion of the total learning task. Far from being just one small 

;. set of signs, with their own peculiar structure and habits, among 

the total inventory of the braille code, these units and shapes 

furnish a predictable feature of nearly every line of braille 

encountered by a fluent reader. Any successful program of teaching 

' must address itself to this task. 

Even a superficial glance at the relative frequencies of the 
lower-sign shapes (contractions) gives us a strong indication of 
■ important parameters. (in) and (to) must clearly be made priority 
items in the braille learning and mastery task. Note that the 
importance of (in) (and (he)) is greatly enhanced by participation 
in the constitution of larger words; the role of (in) as a 



144 



productive prefix shape in English further enhances its frequency 
and salience. The great dominance of (was) over (were) is notable; 
this is easily explained by the fact that in English the plural is 
the marked term in relation to the singular, and therefore may be 
expected to be, as a complex rather than a simple term, less 
frequent than the singular. In all this list (be), (his), (by), 
may be regarded as the median items, for which no special lesson 
may immediately be drawn. 

(into) is in a class by itself, and as a complex in its 
morphology may be expected, like (were) , to show a depressed 
frequency. Relevant experts and users might wish to consider 
redefining (into) as a sequence of its components, and thereby 
remove it as an apparent additional item in this inventory of 
special behavior. 

(enough) is clearly the least urgent task among all of these 
in respect of priority for mastery. Responsible experts might 
profitably consider whether this contraction really pays its way in 
the braille code. Is its yield worth the investment of learning? 



145 



Notes 

^See English Braille American Edition 1959 (Louisville: American 
Printing House for the Blind, 1970) Rule XIII and S39; S. C. 
Ashcroft (1960); E. J. Rex (1970); Lorimer, Tobin, Gill, & Donee 
(1982). 

It is useful to make clear at the outset that in this 
discussion only those lower-sign braille units that represent whole 
English words (i.e. which include logograms with downshifted 
alphabetic shapes) are analyzed and studied. That is to say lower- 
sign non-words (i.e. morphograms and phonogreims with downshifted 
alphabetic shapes, e.g. dis-) are not included in this study. Of 
course, it is not implied that the latter are without interest and 
importance, even relation to our present subject and class. They 
are simply a topic for another day, and should not distract and 
delay us from the present self-evident subject. 

^It is in this spirit and on this basis that Caton, Pester, and 
Bradley (1980) was constructed. As our studies continue and 
findings are refined we must hope to revise that work and other 
like instructional materials, and to reduce their inadequacies. 
^American Association of Workers for the Blind (1913); Report of 
the Uniform Type Committee. The Outlook for the Blind, 7 (1913): 
1-48. Ashcroft (1960); Nolan and Kederis (1969, p. 27-28, 37, 47, 
65-66; p. 87-94 concentrates on upper-dot cells. 

^In the terms of Haimp and Caton (1984) represented as a logogram . 
^In the terms of Hsonp and Caton (1984) represented in phonograms . 
^In the terms of Hamp and Caton (1984) represented in braille 



146 



alphabetic letters . 

'Either as phonograms , morphoqrams , or not, in the terms of Hamp 

and Caton (1984) . 

^That is, the component sequence may appear with its loqoqraphic 

value or simply with the value of a phonogreun corresponding to a 

print sequence. 



;jc; 



147 



References 

American Association of Workers for the Blind. (1913). Report of 

the Uniform Type Committee, Outlook for the Blind, 7, 2-48. 
Ashcroft, S. C. (1960). Errors in oral reading of braille at the 

elementary grade levels. PhD Dissertation, University of 

Illinois, Champaign-Urbana, IL. 
Caton, H., Pester, E., & Bradley, E. J. (1980). Patterns; The 

primary braille reading program. Louisville, KY: American 

Printing House for the Blind 
English braille American edition 1959. (1972). Louisville, KY: 

American Printing House for the Blind. 
Hamp, E. P., & Caton, H. R. (1984). A fresh look at the sign 

system of the braille code. Journal of Visual Impairment & 

Blindness, 78, 210-214. 
KuCera, H., & Francis, W. P. (1967). Computational analysis 

present-day American English. Providence, RI: Brown 

University Press. 
Lorimer, J., Gill, J., Tobin, M. J., & Douce, J. L. (1982). A 

study of braille contractions. London: Royal National 

Institute for the Blind. 
Nolan, C. Y. , & Kederis, C. J. (1969). Perceptual factors in 

braille word recognition. New York: American Foundation for 

the Blind. 



148 



Rex, E. J. (1970). A study of basal readers and experimental 
supplementary instructional materials for teaching primary 
reading in braille. PhD Dissertation, Peabody College of 
Vanderbilt University, Nashville, TN. 

Thorndike, E. L., & Lorge, I. (1944). The teacher' s word book 
of 30,000 words (a number of reprints) . New York: Bureau of 
Publications, Teachers College, Columbia University. 



149 



Suinmary 
Study I. A Fresh Look at the Sign System of the Braille Code 

The internal analysis as described in the 1984 paper 
aimed at devising new descriptions of the elements of the 
braille code. This was done to eliminate the many conflicting 
and confusing terms and categories previously used by teachers 
and students, and to provide them with a new system consisting 
of a relatively small number of categories, or groups, and 
clear, precise, linguistically based descriptions of all the 
elements of English braille. Grade 2. 

In addition, the discussion is intended to clarify the 
internal characteristics of the code. 

We believe that the results of this analysis can be used 



to; 



1. Provide teachers with an outline of braille terms 
that will enable them to describe and discuss any 
element of the braille code in a manner easily 
understood by children. 

2 . Emphasize that the teaching of braille reading and 
print reading are not analogous, and promote an 
understanding of the internal characteristics of 
the braille code, thus providing teachers with more 
effective strategies for teaching reading to 
children who use braille as their primary medium. 

3. Form a basis for any further analysis or productive 
understanding of the function built into the 
structure of the English braille code. 



150 



study II. Toward A Refinement of the Linguistic Analysis of 
American Literary Braille, Grade 2 

1. A plan is outlined to continue and refine the 1984 
analysis, and has been pursued since 1989. 

2 . The conceptual foundations of the analysis are placed in 
the context of 20th century linguistic theory in brief 
and summary form. 

3 . To fulfil the first requisite a suitable text corpus must 
serve as an empirical data base. The creation of such a 
corpus extracted from the "Brown Corpus" is described. 

4. As a sample result, classed totals of braille units (cf . 
the first study) contained in the sample corpus are 
presented. 

Study III. An Analysis of Braille Word Lengths 

1. The lengths of all words in our sample of the Brown 
Corpus (cf. second study above), with the exception of 
certain categories of text spans (cf. Introduction), were 
counted in terms of print characters, braille units, and 
braille shapes, each word being given a characterization 
coding these three counts . The frequency of each of 
these codings and their constituents could then be 
counted. 

2. Short words are most frequent, but more so in braille 
than in inkprint; 63% of braille, and 43% of print words, 
have three or fewer units , or letters . 

3. Nearly 20% of all words contain three print characters. 



151 



yet but one braille shape. 

4. Words with no braille contractions comprise ca. one- 
fourth of the total. 

5. More than one-half of the words show one or two fewer 
shapes in braille than characters in print. 

6. Our computations confirm the saving in lengths through 
braille contractions, but also indicate a general 
correspondence between braille and print lengths and a 
close correspondence between braille unit and shape 
counts. Tabulations bring out deviations and 
correspondences of detail. 

7 . These computations permit the correction of exaggerated 
claims or impressions. 

8 . Accurate data on English word lengths yield implications 
for learning and education and for our understanding and 
criticism of the design of braille. 

9. For a language such as English our ability to supply this 
measure means a clue to a useful metric of English 
discourse and genre. 

Study IV. The Text Frequency and Incidence of Lower-Sign 
Sequences in American Braille 
1. The data for study of the nine lower-sign sequences have 
been taken exhaustively from our sample of the Brown 
Corpus (cf . second study, above) and from a recent number 
of the i^eader's Digest. Both logogram (contraction) and 
spelt-out representations have been studied, and these 



152 



occurring as inkprint words or enclosed in larger words 
or other logograms . Eighteen categories of 
representation are discriminated. All counts are 
subdivided by Brown Corpus genres; this gives the 
possibility of study by discourse and stylistic type. 

2 . Observations are then made on the numerical properties of 
these counts, and visual line graphs are presented. 

3. Some general results emerge for priorities in teaching 
and for relative quantity of work contributed to braille 
text among these lower-sign words. The Reader' s Digest 
sample provides a control on the Brown Corpus count for 
overall incidence. 

4. The decrement in incidence of these words is also 
compared with that found in Lorge-Thorndike 1944. 

5. The question of the importance or salience of these words 
is both broached here and reserved for the future. 



153 



!,=»• 



154 



155 



f' 



156